Carbon Dioxide as Chemical Feedstock

Edited by Michele Aresta Carbon Dioxide as Chemical Feedstock Edited by Michele Aresta Carbon Dioxide as Chemical Fe...

Author: Michele Aresta

235 downloads 2138 Views 5MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Edited by Michele Aresta

Carbon Dioxide as Chemical Feedstock

Edited by Michele Aresta Carbon Dioxide as Chemical Feedstock

Related Titles Paul T. Anastas(Series Editor), Robert H. Crabtree (Editor)

Handbook of Green Chemistry – Green Catalysis 2009 ISBN: 978-3-527-31577-2

Kemmere, M. F., Meyer, T. (eds.)

Supercritical Carbon Dioxide in Polymer Reaction Engineering 2005 ISBN: 978-3-527-31092-0

Horvath, I. T. (ed.) Sheldon, R. A., Arends, I., Hanefeld, U.

Green Chemistry and Catalysis 2007 ISBN: 978-3-527-30715-9

Encyclopedia of Catalysis 6 Volume Set 2003 ISBN: 978-0-471-24183-6

Gabriele Centi, Ferruccio Triﬁró, Siglinda Perathoner, Fabrizio Cavani (eds.)

George A. Olah, Alain Goeppert, G. K. Surya Prakash

Sustainable Industrial Chemistry

Beyond Oil and Gas: The Methanol Economy

2009 ISBN: 978-3-527-31552-9

2006 ISBN: 978-3-527-31275-7

Cornils, B., Herrmann, W. A., Muhler, M., Wong, C.-H. (eds.)

Gadi Rothenberg

Catalysis from A to Z A Concise Encyclopedia 2007 ISBN: 978-3-527-31438-6

Catalysis: Concepts and Green Applications 2008 ISBN: 978-3-527-31824-7

Edited by Michele Aresta

Carbon Dioxide as Chemical Feedstock

The Editor Prof. Dr. Michele Aresta University of Bari Department of Chemistry and CIRCC Via Celso Ulpiani 27 70126 Bari Italy

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliograﬁe; detailed bibliographic data are available on the Internet at . © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microﬁlm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not speciﬁcally marked as such, are not to be considered unprotected by law. Composition Toppan Best-set Premedia Limited, Hong Kong Printing and Bookbinding betz-druck GmbH, Darmstadt Cover Design Schulz Graﬁk-Design, Fußgönheim Printed in the Federal Republic of Germany Printed on acid-free paper ISBN: 978-3-527-32475-0

V

To Federica, Mattia and Nicolò

Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

VII

Contents Preface XV List of Contributors 1

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8

2

2.1 2.2 2.3 2.4 2.5 2.5.1 2.5.2 2.6

XVII

Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere 1 Michele Aresta Carbon Dioxide Emission 1 The Accumulation of CO2 in the Atmosphere, and the Effects that We Fear 1 Technologies to Reduce CO2 Accumulation in the Atmosphere 4 The Utilization of CO2 6 Conditions for Using CO2 8 CO2: Sources and Prices 8 The Potential for CO2 Utilization, and the Content of This Book 9 The Need for Research to Speed an Exploitation of the Utilization Option 11 References 13 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses 15 Alessandro Galia and Giuseppe Filardo Introduction 15 Dense Carbon Dioxide as Solvent Medium for Chemical Processes 15 Enzymatic Catalysis in Dense Carbon Dioxide 18 Other Reactions in Dense Carbon Dioxide 19 Polymer Synthesis in Supercritical Carbon Dioxide 20 Chain Polymerizations: Synthesis of Fluoropolymers 22 Step Polymerizations: Synthesis of Biodegradable Polymers Conclusions 27 Acknowledgments 27 References 28

Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

26

VIII

Contents

3

3.1 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 3.2.6 3.3 3.4 3.5

4

4.1 4.2 4.3 4.3.1 4.3.2 4.3.2.1 4.3.2.2 4.3.2.3 4.3.2.4 4.4 4.4.1 4.4.2 4.4.3 4.5 4.5.1 4.5.2 4.5.3 4.5.4 4.6

Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations 33 Ivan A. Berg, Daniel Kockelkorn, W. Hugo Ramos-Vera, Rafael Say, Jan Zarzycki, and Georg Fuchs Introduction 33 The Mechanisms of CO2 Fixation 34 The Calvin–Benson–Bassham (CBB) Cycle 34 The Reductive Citric Acid Cycle (Arnon–Buchanan Cycle) 37 The Reductive Acetyl-CoA Pathway (Wood–Ljungdahl Pathway) 39 The 3-Hydroxypropionate/Malyl-CoA Cycle 40 The 3-Hydroxypropionate/4-Hydroxybutyrate Cycle 42 The Dicarboxylate/4-Hydroxybutyrate Cycle 44 Rules to Explain the Diversity 46 Evolutionary Aspects 49 Chemical Aspects of CO2 Fixation 50 Acknowledgments 51 References 51 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2 55 Joëlle Mascetti Introduction 55 Carbon Dioxide Bonding to Metals 56 Synthesis and Structure of CO2 Complexes 59 Low-Temperature Matrix Isolation and Theoretical Studies 59 Synthesis of Stable Complexes 64 End-On Complexes 65 Side-On Complexes 67 Bridged Complexes 67 Bridged Complexes Obtained by In-situ Synthesis 67 Reactivity of CO2 Complexes 69 C–O Bond Cleavage and O Transfer 70 Reactions with Electrophiles 72 Reactions with Nucleophiles 73 CO2 Complexes as Reaction Intermediates in CO2 Utilization Processes 75 Oxidative Coupling Reactions 76 Reduction Reactions 79 Catalytic Processes 81 Bioinspired Reactions 82 Conclusions 84 Acknowledgments 85 References 85

Contents

5

5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.3.1 5.3.2 5.3.2.1 5.3.2.2 5.3.2.3 5.4

6

6.1 6.2 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.3.6 6.4 6.5 6.6

7 7.1 7.2

Main Group Element- and Transition Metal-Promoted Carboxylation of Organic Substrates (Alkanes, Alkenes, Alkynes, Aromatics, and Others) 89 Thomas Zevaco and Eckhard Dinjus Introduction 89 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis 90 Kolbe–Schmitt Synthesis: Generalities 90 Reaction Parameters and Mechanistic Studies of the Kolbe–Schmitt Synthesis 91 Recent Applications of the Kolbe–Schmitt Carboxylation: Synthesis of 1,3-Dialkylimidazolium-2-Carboxylates 97 Carboxylation of C–H-Acidic Compounds 99 Reactive Organometallic Derivatives in the Synthesis of Carboxylic Acids 102 Generalities 102 Bimetallic Catalytic Systems 104 Pd/Sn Systems 104 Rh/B and Cu/B 107 Ni/Zn 108 Palladium (0)-Catalyzed Telomerization of Butadiene with CO2: Synthesis of δ-Lactone 112 References 116 The Chemistry of N–CO2 Bonds: Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas 121 Eugenio Quaranta and Michele Aresta Introduction 121 Synthesis of Carbamic Acids and Alkylammonium Carbamates 122 Synthesis of Carbamate Esters 125 Transfer of Carbamate Group to Alkyl Halides 126 Transfer of Carbamate Group to Acylating Agents 131 Transfer of Carbamate Group to Alcohols 134 Transfer of Carbamate Group to Epoxides 138 Transfer of Carbamate Group to C–C Double Bonds 142 Transfer of Carbamate Group to C–C Triple Bonds 145 Synthesis of Isocyanates 148 Synthesis of Ureas 154 Conclusions 159 References 159 Synthesis of Linear and Cyclic Carbonates 169 Danielle Ballivet-Tkatchenko and Angela Dibenedetto Introduction 169 Acyclic Organic Carbonates 170

IX

X

Contents

7.2.1 7.2.2 7.2.3 7.2.4 7.2.5 7.3 7.3.1 7.3.1.1 7.3.1.2 7.3.1.3 7.3.1.4 7.3.2 7.3.2.1 7.3.2.2 7.3.3 7.3.3.1 7.3.3.2 7.3.3.3 7.3.4 7.3.5 7.3.6 7.3.7 7.3.8 7.4 7.4.1 7.4.2

8

8.1 8.2 8.3 8.4 8.4.1 8.4.2 8.5 8.5.1 8.5.2 8.5.3 8.6

Market and Production 170 Current Trends with CO2 as Feedstock 172 Alcoholysis of Urea 174 Direct Route 175 The Future of CO2-Based Routes to Acyclic Carbonates 180 Synthesis of Organic Cyclic Carbonates 181 Carboxylation of Epoxides 182 Use of Conventional Solvents 182 Use of Ionic Liquids 184 Use of Supercritical CO2 185 Combined Reaction Media: sc-CO2 and Ionic Liquids 185 Oxidative Carboxylation of Oleﬁns 186 Use of Oxygen as Oxidant 186 Use of Other Oxidants 188 Other Synthetic Routes to Cyclic Carbonates 189 From Halohydrins 189 From Halogenated Carbonates 190 Reaction of Cyclic Ketals with Carbon Dioxide 191 Synthesis of Cyclic Carbonates from Propargylic Alcohols 191 Reaction Between Carbon Dioxide and Diols 192 Reaction of Urea and Diols 194 Reaction of Carbon Dioxide or Urea with Glycerol 195 Reactivity of Cyclic Alkylene Carbonates 198 Transesteriﬁcation Reactions 200 Synthesis of Acyclic Carbonates 200 Synthesis of Cyclic Carbonates 204 References 205 Polymers from Carbon Dioxide: Polycarbonates, Polythiocarbonates, and Polyurethanes 213 Donald J. Darensbourg, Jeremy R. Andreatta, and Adriana I. Moncada Introduction 213 Historical Perspective 215 Metal Catalysts for the Copolymerization of Epoxides and CO2 215 Metal Catalysts for the Copolymerization of Oxetanes and CO2 228 Early Studies 228 Recent Studies using Transition Metal Catalysts 229 Physical Methods for the Characterization of Copolymers Produced from Epoxides or Oxetane and Carbon Dioxide 235 Cyclohexene Oxide Monomer 235 Propylene Oxide Monomer 237 Oxetane Monomer 239 Copolymer Isolation and Catalyst Recycling 241

Contents

8.7 8.8 8.9

9

9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9

10

10.1 10.2 10.2.1 10.2.2 10.2.3 10.3 10.3.1 10.3.2 10.3.3 10.3.4 10.3.5 10.4 10.4.1 10.4.2

Copolymerization of Carbon Disulﬁde and Epoxides and Episulﬁdes 244 Copolymers from Aziridines and Carbon Dioxide 245 Concluding Remarks 245 Acknowledgments 246 References 246 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane for Synthesis Gas Production, Using the Tapered Element Oscillation Microbalance 249 Wie Pan and Chunshan Song Introduction 249 Thermodynamic Analysis of Carbon Formation from CH4 or CO 252 Thermodynamic Analysis of Carbon Formation in CO2 Reforming of Methane 254 TEOM Measurement of Carbon Formation in CO2 Reforming of Methane 256 TPO Analysis of Carbon Formation in CO2 Reforming 257 TEM Analysis on Carbon Formed on Catalysts After CO2 Reforming of Methane 259 Kinetic Study of Carbon Formation on PCH4 and PCO in CO2 Reforming 260 H2O Effect on Carbon Formation in CO2 Reforming 262 Conclusions 263 Acknowledgments 263 References 263 Appendix A9.1 264 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches 267 Ji-Jun Zou and Chang-Jun Liu Introduction 267 Nonthermal Plasma Phenomena 268 Electron/Molecular Reactions 270 Atom/Ion/Molecule Reactions 270 Heterogeneous Reactions 271 CO and/or H2 Production from CO2 272 CO2 Dissociation 272 Reforming of CH4 with CO2 273 Reforming of Aliphatic Hydrocarbons with CO2 275 Other Reforming Reactions with CO2 276 Reduction of CO2 276 Hydrocarbons Synthesis from CO2 277 Oxidative Coupling of CH4 with CO2 277 Hydrogenation of CO2/CO to C2 Hydrocarbons 278

XI

XII

Contents

10.4.3 10.5 10.5.1 10.5.2 10.5.3 10.5.4 10.5.5 10.6 10.6.1 10.6.2 10.7

Higher Hydrocarbons from CH4 and CO2 279 Oxygenates Synthesis from CO2 280 Methanol from CO2 Hydrogenation 280 Methanol from CO2 and CH4 281 Aldehydes from CH4 and CO2 282 Acetic Acid from CH4 and CO2 282 Oxygenates from H2O and CO2 283 Combination of Plasma with Catalyst 284 Catalysts in Plasma Utilization of CO2 284 Interaction Between Plasma and Catalyst 285 Summary 286 Acknowledgments 287 References 287

11

Photochemical, Electrochemical, and Photoelectrochemical Reduction of Carbon Dioxide 291 Emily Barton Cole and Andrew B. Bocarsly Introduction 291 Homogeneous Photochemical Reduction 292 General Considerations 292 Transition Metal Complexes 294 Ruthenium Complexes 294 Rhenium Complexes 295 Macrocyclic Complexes 296 Electrochemical Reduction 296 Reduction in Aqueous Solutions at Metal Electrodes 297 Reduction to CO and HCOOH 298 Reduction to CH3OH and Alcohols 299 Reduction to CH4 and Hydrocarbons 300 Reduction in Nonaqueous Solutions at Metal Electrodes 302 Reduction Mediated by Metal Complexes 303 Transition Metal Complexes 303 Macrocyclic Complexes 304 Metal-Containing Enzyme-Mediating Complexes 305 Semiconductor Systems for Reduction 305 Photoelectrochemical Semiconductor Electrode Systems 305 Unmodiﬁed Semiconductor Electrode Interfaces 306 Modiﬁed Semiconductor Electrodes 307 Homogenous Solution Catalysts at Semiconductor Electrodes 308 Heterogeneous Photochemical Semiconductor Systems 309 Unmodiﬁed Semiconductor Colloids and Powders 309 Metal-Coated Semiconductor Colloids and Powders 310 Concluding Remarks and Future Directions 311 References 312

11.1 11.2 11.2.1 11.2.2 11.2.2.1 11.2.2.2 11.2.3 11.3 11.3.1 11.3.1.1 11.3.1.2 11.3.1.3 11.3.2 11.3.3 11.3.3.1 11.3.3.2 11.3.3.3 11.4 11.4.1 11.4.1.1 11.4.1.2 11.4.1.3 11.4.2 11.4.2.1 11.4.2.2 11.5

Contents

12

12.1 12.2 12.2.1 12.2.2 12.2.3 12.3

13

13.1 13.2 13.3 13.3.1 13.3.2 13.4 13.5 13.5.1 13.5.2 13.6 13.7

14

14.1 14.2 14.3 14.3.1 14.3.2 14.3.3 14.3.3.1 14.3.3.2 14.3.3.3 14.3.3.4 14.4 14.4.1 14.4.2

Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation Based on Carbon Dioxide 317 Giuseppe Silvestri and Onofrio Scialdone Introduction 317 Electrocarboxylation 318 Electrocarboxylation of Organic Halides 319 Electrocarboxylation of Aromatic Ketones 324 Electrocarboxylation of Other Substrates 326 The Electroreduction of Carbon Dioxide in Protic Media (Water and Alcohols) 327 Acknowledgments 330 References 330 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass 335 Michele Aresta and Angela Dibenedetto Introduction 335 The Natural Carbon Cycle 336 The Utilization of Terrestrial Biomass 337 Residual Biomass 338 Cultivated Biomass 339 The First-Generation Biofuels 339 The New Generations of Biofuels 339 Second-Generation Biofuels 340 Third-Generation Biofuels 341 Implementation of the Bioreﬁnery Concept 347 Concluding Remarks 349 References 349 Fixation of Carbon Dioxide into Inorganic Carbonates: The Natural and Artiﬁcial “Weathering of Silicates” 353 Ron Zevenhoven and Johan Fagerlund Introduction: Inorganic Carbonate Uses and Natural Resources 353 Natural Fixation of CO2 in Carbonates 355 Process Routes to Valuable Carbonate Products 357 Material Resources 357 Direct (Single-Step) Process Routes 358 Indirect (Single-Step) Process Routes 359 General Aspects of Calcium Carbonate Production 359 Acetic Acid Route 361 Two-Step Aqueous Carbonation of Solid Residues 362 The pH-Swing Process 363 Mineral Carbonation for Carbon Capture and Storage (CCS) 364 Material Resources 366 Direct (Single-Step) Process Routes 367

XIII

XIV

Contents

14.4.2.1 14.4.2.2 14.4.3 14.4.3.1 14.4.3.2 14.5 14.5.1 14.5.2

Gas–Solid Processes 367 Aqueous Solution Processes 369 Indirect (Multistep) Process Routes 369 Gas–Solid Processes 369 Aqueous Solution Processes 372 Other Carbonate Production Processes and Applications Carbonation of Brines 374 Straightforward Carbonation 375 Acknowledgments 375 References 376 Index 381

374

XV

Preface

In Nature, the carbon cycle has the ability to recycle some 203 gigatons (Gt) of carbon dioxide each year (see Figure 13.1). Although anthropogenic (“man-made”) CO2 which, at 7 Gt per year, represents approximately 3.4% of the total CO2 converted in the natural cycle – an apparently small amount – the natural carbon cycle, despite having become extremely proﬁcient over millions of years, is unable to recycle this excess CO2. As a result, CO2 becomes accumulated in the atmosphere, leading in turn to an expansion of the “greenhouse effect” with, potentially, a major impact on climate change. Hence, it is vital that strategies are developed to limit CO2 accumulation in the atmosphere. The use of CO2 – as either a technological ﬂuid or a raw material in chemical processes and in biotechnological applications, such as enhanced biological ﬁxation – provides the potential to reduce CO2 emissions. Indeed, CO2 utilization may become an efﬁcient tool if it can be merged with the development of innovative, “sustainable” technologies with less-intensive energy and materials requirements than those currently on stream. For this purpose, CO2 may be used either as a technological ﬂuid, as a building block for complex molecules, or as a carbon source for fuels. It is important to remember that, whichever compound is prepared from CO2, during the use of that compound the CO2 will be returned. The competition is, therefore, between the rate of formation of CO2 and the rate of its ﬁxation. Most CO2 is produced in combustion processes which, in general, are very fast – much faster than any reaction that may convert CO2 into chemicals. Even natural processes that ﬁx CO2 into biomasses function much more slowly than combustion. In fact, the rate of depletion of natural fossil sources is higher than the rate of formation of fossil carbon, and consequently any competition based on the rapidity of these opposing processes will be lost in favor of CO2 production. Thus, it is unrealistic to imagine that the conversion of CO2 might solve the problem of its accumulation in the atmosphere, as kinetic – and often also thermodynamic – factors are braced against such possibility. Hopefully, it might in time be possible to convert a limited fraction – perhaps up to 7% – of the anthropogenic CO2 back into useful chemicals or fuels. Although this will not solve the problem of CO2 accumulation in the atmosphere, when added to other CO2 utilization technologies it Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

XVI

Preface

may provide an effective level. In fact, no technology is available today that is capable of performing much better under the same conditions of energy consumption and economics. The use of CO2 represents one option that might provide a serious contribution to reducing CO2 accumulation in the atmosphere, speaking not in terms of its storage but rather but in terms of its efﬁciency. An exception might be to use CO2 as a co-monomer so as to afford polycarbonates and polyurethanes that may persist for decades and, therefore, represent an optional means of CO2 storage. Carbon dioxide is considered to be an inert molecule since, with water, it is the end product of any combustion process, including biological cellular oxidation reactions. Although it is produced by all living organisms, whether animal or vegetable (for example, an adult man emits about 0.9 kg CO2 per day), by far the main source of CO2 is the combustion of fossil carbon (coal, oil, gas) used for the production of energy. This book contains highlights of the research carried out during the past 20 to 30 years or so, with emphasis placed not only on fundamental concepts but also on applications to synthetic chemistry and the development of new processes based on metal-catalyzed conversion reactions. The collection of review material gathered here provides a cross-disciplinary insight into a very wide-ranging ﬁeld that bridges the chemical and biotechnological aspects of CO2 conversion with its technological use. The most recent discoveries and creative synthetic strategies are discussed in detail, while the barriers that must be overcome in order to fully exploit the utilization option are highlighted. This book represents an updated tool for those teachers, research workers and students who wish to become acquainted with the most advanced knowledge of CO2 chemistry, and strategies for its conversion and utilization. Consequently, these chapters offer a diversity of options in such a complex area, and illustrate the links between chemistry and biology, showing how hybrid technologies may contribute to identify operative and effective solutions to these problems. The challenges within the different areas are launched through open questions, and it is hoped that the readers may indeed use their own initiative to discover new and more efﬁcient solutions. In this way, a practical application may be found that will transform a dream into reality. I wish to express my most sincere thanks to my colleague Professor Angela Dibenedetto for her precious collaboration, and her excellent help provided during the preparation of this book. University of Bari Department of Chemistry and Interuniversity Consortium on Chemical Reactivity and Catalysis-CIRCC Campus Universitario 70126 Bari, Italy June 2009

Michele Aresta

XVII

List of Contributors Jeremy R. Andreatta Texas A&M University Department of Chemistry College Station, TX 77843 USA

Andrew B. Bocarsly Princeton University Frick Laboratory Princeton, NJ 08544 USA

Michele Aresta University of Bari Department of Chemistry and CIRCC 70126 Bari Italy

Donald J. Darensbourg Texas A&M University Department of Chemistry College Station, TX 77843 USA

Danielle Ballivet-Tkatchenko Université de Bourgogne and CNRS, UMR 5260 Institut de Chimie Moléculaire 9, avenue A. Savary 21000 Dijon France

Angela Dibenedetto University of Bari Department of Chemistry and CIRCC Campus Universitario 70126 Bari Italy

Emily Barton Cole Princeton University Frick Laboratory Princeton, NJ 08544 USA Ivan A. Berg Universität Freiburg Fakultät Biologie Schänzlestr. 1 79104 Freiburg Germany

Eckhard Dinjus Forschungszentrum Karlsruhe GmbH Institute of Technical Chemistry Division of Chemical–Physical Processing Hermann-von-Helmholtz-Platz 1 76344 Eggenstein-Leopoldshafen Germany Johan Fagerlund Abo Akademi University 20500 Abo/Turku Finland

Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

XVIII

List of Contributors

Giuseppe Filardo Università di Palermo Dipartimento di Ingegneria Chimica dei Processi e dei Materiali Viale delle Scienze Ed. 6 90128 Palermo Italy Georg Fuchs Universität Freiburg Fakultät Biologie Schänzlestr. 1 79104 Freiburg Germany Alessandro Galia Università di Palermo Dipartimento di Ingegneria Chimica dei Processi e dei Materiali Viale delle Scienze Ed. 6 90128 Palermo Italy Daniel Kockelkorn Universität Freiburg Fakultät Biologie Schänzlestr. 1 79104 Freiburg Germany Chang-Jun Liu Tianjin University Key Laboratory for Green Chemical Technology of Ministry of Education School of Chemical Engineering and Technology Tianjin, 300072 China Joëlle Mascetti Université Bordeaux 1 Institut des Sciences Moléculaires (UMR 5255 CNRS) 351 cours de la Libération 33405 Talence Cedex France

Adriana I. Moncada Texas A&M University Department of Chemistry College Station, TX 77843 USA Wie Pan Pennsylvania State University Clean Fuels and Catalysis Program EMS Energy Institute, and Department of Energy & Mineral Engineering 209 Academic Projects Building University Park, PA 16802 USA Eugenio Quaranta University of Bari Department of Chemistry and CIRCC 70126 Bari Italy W. Hugo Ramos-Vera Universität Freiburg Fakultät Biologie Schänzlestr. 1 79104 Freiburg Germany Rafael Say Universität Freiburg Fakultät Biologie Schänzlestr. 1 79104 Freiburg Germany Onofrio Scialdone University of Palermo Dipartimento di Ingegneria Chimica dei Processi e dei Materiali Viale delle Scienze 90128 Palermo Italy

List of Contributors

Giuseppe Silvestri University of Palermo Dipartimento di Ingegneria Chimica dei Processi e dei Materiali Viale delle Scienze 90128 Palermo Italy Chunshan Song Pennsylvania State University Clean Fuels and Catalysis Program EMS Energy Institute, and Department of Energy & Mineral Engineering 209 Academic Projects Building University Park, PA 16802 USA Jan Zarzycki Universität Freiburg Fakultät Biologie Schänzlestr. 1 79104 Freiburg Germany

Thomas Zevaco Forschungszentrum Karlsruhe GmbH Institute of Technical Chemistry Division of Chemical–Physical Processing Hermann-von-Helmholtz-Platz 1 76344 Eggenstein-Leopoldshafen Germany Ron Zevenhoven Abo Akademi University 20500 Abo/Turku Finland Ji-Jun Zou Tianjin University Key Laboratory for Green Chemical Technology of Ministry of Education School of Chemical Engineering and Technology Tianjin, 300072 China

XIX

1

1 Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere Michele Aresta

1.1 Carbon Dioxide Emission

Carbon dioxide (CO2) is considered to be the major cause of climate change, because of its greenhouse properties and continuous accumulation in the atmosphere. Indeed, the atmospheric concentration of CO2 has risen from 278 ppm during the preindustrial era to a current level of 387 ppm (Figure 1.1) [1]. The origin of this massive emission of CO2, and of its steady accumulation in the atmosphere, has been the use of carbon-based fossil fuels in human activities (Figure 1.2). Moreover, as carbon-based fossil fuels currently represent 80–85% of the world’s energy sources, and will continue to play such role at least in the short to medium term, a major expansion is visualized in the emission of CO2 (Figure 1.3). Today, the robust growth of emerging economies such as China and India is driving worldwide energy demand and usage such that is increasing at a rate never before experienced; in fact, a further expansion of 50–100% is envisaged by the year 2030 [2]. The novel point here is that, compared to the recent past, these emerging economies will drive the expansion [2] of energy use, with the energy requirements of the OECD (Organisation for Economic Co-operation and Development) countries envisaged to grow at a rate of 0.7% per year, while that of nonOECD countries will grow at 2.5% per year. For example, China and India increased their combined share of world energy consumption from 8% in 1980, to 18% in 2005. Moreover, it is estimated that, in 2009 the energy consumption of China will become equal to that of the USA, and that by the year 2010 China will become the world leader in terms of CO2 emission.

1.2 The Accumulation of CO2 in the Atmosphere, and the Effects that We Fear

Today, as CO2 continues to accumulate in the atmosphere, serious concerns are being raised regarding its inﬂuence on climate change. The greenhouse effect that is associated with energy production and use – whether by the direct release into Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

1 Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere

Figure 1.1 The steady increase in CO2 concentrations in the atmosphere in recent years. Solid and open symbols indicate average and seasonal trends, respectively.

8000 7000 6000

Total CO2 emissions from fossil-fuels (million metric tons of C) CO2 emissions from gas fuel consumption CO2 emissions from liquid fuel consumption CO2 emissions from solid fuel consumption CO2 emissions from cement production CO2 emissions from gas flaring

Million metric tons of carbon

2

5000 4000 3000 2000 1000 0 1751

Figure 1.2

1801

1851

1901

1951

2001

The origin of the emission of CO2.

the atmosphere of ﬂue gas-associated heat produced by the energy, industrial, and transport sectors, or by the emission of greenhouse gases – is slowly causing the temperature of the Earth to rise, as revealed in a report by the Intergovernmental Panel on Climate Change (IPCC) (Figure 1.4) [3]. Inevitably, these changes in the thermal structure of the atmosphere, and of the Earth’s solid surface, are causing the temperatures of the planet’s waters to rise and their volume to increase, to a point where the rising water levels on the Earth’s surface are seen as a major problem, especially in coastal areas.

1.2 The Accumulation of CO2 in the Atmosphere, and the Effects that We Fear 40.000 35.000 30.000

3

mongabay.com United States Mexico South Korea Russia India Africa OECD Europe

Canada Japan Australia/New Zealand China Middle East Brazil source data: EIA 2007

OECD Europe Brazil Africa Middle East India

25.000 20.000

China

15.000

Russia Australia/NZ South Korea Japan Mexico Canada United States

10.000 5.000 0 1990

2002

2003

2010

2015

2020

2025

Figure 1.3 Scenario of the emission of CO2 until 2030. In 2010, China will become the leading CO2 emitter worldwide.

Figure 1.4 Variation of the temperature of the Earth during the last millennium. The upper panel shows the trend during the past 140 years (IPCC data).

2030

4

1 Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere

Figure 1.5 Correlation among the emission of CO2 into the atmosphere, the time of stabilization of its concentration, the temperature variation, and the rise in sea-level.

Unfortunately, the time required to stabilize the effects of increased greenhouse gas concentrations in the atmosphere will be very long. Any response by the Earth to measures that we might take today will not be effective tomorrow, nor the day after tomorrow – it will take centuries (Figure 1.5). So, with such a long time-scale, the ability to forecast scenarios will be affected by any major errors that might emerge as a result of any, often unpredictable, changes that might occur. During the past twenty years, there have been major changes in the social, political, and economical structures of a large part of society that, in 1985, were not considered realistic when forecast scenarios were ﬁrst constructed. The subsequent development of a market economy in Eastern Europe and Russia has led to large demands for energy, whilst the current changes in China and India are creating much uncertainty with regards to future energy needs. Today, the economies of large areas of the Earth, including Eastern Europe and Russia, Southern America, other Asian countries in addition to China and India, and almost the whole Africa, are undergoing continuous evolution to a point where their real rate of growth is not exactly known. Clearly, under such conditions the variability and uncertainty of any foreseen scenario is very large, and it will inevitably become essential that responsible measures are taken in order to avoid reaching a ‘point of no return.’

1.3 Technologies to Reduce CO2 Accumulation in the Atmosphere

It has been predicted that carbon-based fossil fuels will continue to provide 80– 85% of the world energy consumption at least until 2030. Consequently, it is clear that, despite uncertainties regarding the role of atmospheric CO2 on climate

1.3 Technologies to Reduce CO2 Accumulation in the Atmosphere

change and on the occurrence of catastrophic events, as well as links between population growth and energy consumption, the predicted large rise in CO2 emissions is raising serious concern in relation to the rise in the Earth’s temperature. Today, the need to control CO2 production and emissions is the center of attention of both scientiﬁc and industrial worlds and, as a consequence, several methods have been developed to achieve this goal (Table 1.1). None of the technologies listed in Table 1.1 can alone provide a short- to medium-term solution to reduce atmospheric CO2 emissions at a level necessary to stabilize current concentrations. Rather, the correct blend of technologies

Table 1.1

Technologies for controlling the emission of CO2 into the atmosphere.

Technology

Examples of application

Comments

Efﬁciency

Production of electric energy

The efﬁciency of the technologies of conversion of chemical-into-electric energy may rise from 32% to over 50%

Use of any form of energy

Saving energy through a responsible use

Fuel shift

Substitution of coal with oil or gas

Emission value expressed as kg CO2 per kWh electric energy produced: 1 for coal, 0.75 for oil, 0.5 for gas.

Advanced technologies for electric energy production

IGCC

Concentration of the production of CO2 via decarbonization of fossil fuels. CO2 can be either disposed of or used

Non-carbon-based fuels

Nuclear energy

Use of nuclear fuels for the production of electric energy; used by sectors that require a high intensity

Perennial energies

Solar, wind, hydro, geothermal

Their exploitation depends on the geographic position of a country

Renewables

Biomass utilization

Residual and cultivated terrestrial and aquatic biomass can produce liquid or gaseous fuels suitable for the transportation sectors.

Carbon dioxide capture and storage (CCS)

Capture and disposal of CO2

This technology has a large disposal potential, but sites are not available in all counties. It is energy-intensive and will expand the use of fossil-C well beyond the current limits

CO2 utilization

CO2 is recycled, mimicking Nature

Chemical, technological, and enhanced-biological uses are the only ones that generate a proﬁt, adding value to a waste

5

1 Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere 38

36

36

36

34

34 32

30

30 Volume Energy Density/GJ/m3

6

28 26 24 22 20

18 17

18 16

13

14 12

9

10

8

8 6 4

2

2 0 diesel

Figure 1.6

bio-diesel from algae

gasoline

carbon coke

brown coal

methanol

bio-oil from algae

H2 (l)

methane (g)

H2 (g) 20,0 MPa

Energy density of several different liquid and gaseous vectors.

must be identiﬁed to achieve the most effective reduction in CO2 emissions, while paying the lowest social (by boosting development) and economic (affordable technologies for all societies) costs. In this way, these methods may be developed over time such that they become effective for application on the large scale. Likewise, the optimal use of resources may be maximized by applying intelligent solutions. Based on the different intensities of energy required, industrial and collective electriﬁed transport uses may be separated from personal transportation and municipal uses of energy. For example, gasoline, diesel and methane might represent the best options for the individual and mass road-transportation sector, as they have the correct energy density (Figure 1.6), whereas ‘renewables’ might represent the ideal candidate to partially substitute fossil fuels in the transportation sector and in the chemical industry. Perennial sources could be used to best advantage in the short term by municipalities or small industries, fossil carbon could be used in the production of certain necessary chemicals or fuels, while nuclear power would be better used for industrial other applications that demand a highly intensive use of energy.

1.4 The Utilization of CO2

In 2008, increases in the price of oil and the recognition of a need to reduce the impact of the chemical and energy industries on climate change, notably by reducing the emission of CO2, directed interest towards the use of renewable sources of energy and alternative feedstocks for the chemical industry. This strategy

1.4 The Utilization of CO2 Coal, 23% Natural gas, 24% Biomass, 47% Renewable energy, 6% Nuclear, 8%

Hydroelectric, 45% Geothermal, 5% Wind, 2% Solar, 1%

Petroleum, 39% Figure 1.7

Potential of biomass in the USA (from EIA-Paris, 2004).

included the industrial utilization of CO2 and the enhanced ﬁxation of CO2 in aquatic biomass, with both applications having huge potential for the recycling of carbon and reductions in CO2 emissions. The enhanced, direct ﬁxation of CO2 into fast-growing biomasses might contribute towards reducing its accumulation in the atmosphere, under non-natural conditions. Such an approach could be used for the production of chemicals and energy (e.g., conversion into gaseous and liquid fuels, rather than direct combustion of the solid biomass), with beneﬁcial effects on reducing CO2 emissions and accumulation in the atmosphere. The potential of biomass as a possible substitute for fossil fuels in the USA is shown in Figure 1.7. The implementation of an artiﬁcial photosynthetic cycle which could recycle CO2 [4, 5], thus complementing the natural cycle, could also make an effective contribution to reducing CO2 emissions. The utilization of CO2 through technological, chemical, or enhanced biological methods may lead to reductions in CO2 emissions, with these innovative processes being substituted for older technologies and products, and imparting direct and indirect beneﬁts on the impact of climate change. In this case, a direct effect might be considered the reduction of both CO2 emissions and fossil fuels extractions, whereas an indirect effect would be a substitution, with CO2, of chemicals such as chloroﬂuorocarbons and congeners having a climate change power (CCP) many thousand-fold that of CO2 [6]. It should be remembered that, whilst the use of CO2 in this way will not solve the problem of atmospheric CO2 accumulation, it might contribute to such an issue by reducing the volume of CO2 produced. It is also worth noting that the fraction of CO2 produced via the use of chemicals is approximately 10% of the total, the remainder being derived from energy products. It follows, therefore, that if all chemicals were to be produced from CO2, then 10% of the emitted CO2 could be recycled so as to achieve an almost carbon-neutral use of chemicals. Unfortunately, this situation would be impossible to achieve for several reasons, among which are the energy costs. A realistic estimate of the total amount of CO2 to be avoided [7] has been set at 7%, this being the contribution of CO2 use in the short to medium term. However, if efﬁcient technologies capable of converting CO2 into energy-rich products (fuels) were to be developed, then a much greater amount of

7

8

1 Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere

CO2 could be converted into usable products. This would result in a much more signiﬁcant step in the direction of chemicals and energy production, with a close to zero carbon-emission level.

1.5 Conditions for Using CO2

The utilization of CO2, whether technological, biological, or chemical, to cause an effective reduction in its emission into the atmosphere, must comply with certain key rules, namely:

• • • •

the new process must reduce the overall CO2 emissions; it must be less energy- and material-intensive with respect to the on-stream processes that it aims to replace; it must employ using safer and more eco-friendly working conditions; and it must be economically viable.

The reduction in overall CO2 emission for a given application of CO2 is not easily quantiﬁed. Both, the energy and material consumption must be minimized, through the control of process parameters, such as the conversion yield and selectivity, the temperature, the pressure, and the post-reaction operations such as isolation and puriﬁcation. Mass control requires more direct (fewer steps), effective (high-yield) and selective (product entropy control) processes, with waste (gas, liquid, solid) minimization at source, and with a better carbon-atom utilization. As CO2 is not a toxic substance, under controlled conditions of utilization it does not cause any serious worries. In particular, its ﬁre-extinguishing properties ensure that the risk of combustion is close to zero when a high pressure is used [30–40 MPa when supercritical CO2 (sc-CO2) is used as solvent and/or reagent]. Hence, CO2 can be considered to be safe reagent or solvent, especially when sc-CO2 is used as a process ﬂuid (see Chapter 2). The economics of CO2 utilization depend on its quality, with the price reaching up to US$ 400 per ton, according to the purity required and the puriﬁcation technologies involved [8].

1.6 CO2: Sources and Prices

As noted above, CO2 can be obtained from several sources, with a variety of prices. Although natural wells are an important and cheap source (€15–20 per ton) [9] of pure (>99%) CO2, this source should be discontinued and the use of recovered CO2 (from power-generation plants or industrial processes) encouraged as an alternative. The captured CO2 is characterized by different degrees of purity, according to its origin, and may require extensive puriﬁcation in speciﬁc applications, such as in the food industry. These puriﬁcation steps will, of course, affect

1.7 The Potential for CO2 Utilization, and the Content of This Book

the price of CO2, and this will in turn invite the discovery and use of new sources, an example being fermentation reactors which may provide high-purity gas. One possible drawback of such a source might be its seasonality (e.g., sugar cane harvests), although this could easily be circumvented by the year-round storage of these raw materials. In the case of several applications in the chemical industry, the presence of contaminants such as O2, SOx, or NOy, which may be present in the ﬂue gases of power plants, might be deleterious by causing negative (poisoning) effects on the catalysts.

1.7 The Potential for CO2 Utilization, and the Content of This Book

The large-scale separation of CO2 from power and industrial plant ﬂue gases, or its recovery via new technological applications (e.g., integrated gasiﬁcation combined cycle; IGCC) which produce concentrated ﬂows of CO2, will make available huge volumes of CO2. The subsequent fate of the recovered CO2 would be either disposal in natural ﬁelds (spent gas or oil wells, unmined coal seams, aquifers), or its utilization. Whilst, in principle, the former option can be used to eliminate large volumes of CO2, in practice the disposal sites are few in number and the procedure would involve high energy and economic costs. Eventually, however, the implementation of this technology will surely lead to an expansion of the extraction of fossil carbon-based fuels. Conversely, the utilization of CO2 has already been implemented at the level of over 130 Mt per years in the energy and chemical industries [6], and is a technology that is known to add value to the waste CO2. Today, new technologies using CO2 are under development, the exploitation of which will greatly expand the amount of CO2 that can be either used or recycled. The large-scale utilization of CO2 can be integrated very well with its large-scale recovery, yet, whilst the utilization process will produce economic beneﬁts, the disposal process will incur economic costs. It is essential that these new applications of CO2 are energetically more convenient than the existing processes, and in this respect the biological [10] and technological [6, 11] applications of CO2 utilization have been extensively reviewed. Indeed, the aim of this book is to describe in detail the chemistry of the utilization of CO2, presenting an up-date in each speciﬁc ﬁeld of CO2 use. The technological use of CO2 contributes to a reduction in its atmospheric accumulation, through CO2 being substituted into species with a much higher CCP [6]. Some examples where the use of CO2 lowers the impact on climate change (even if the CO2 is ultimately vented to the atmosphere) are listed in Table 1.2. Among the applications listed in Table 1.2, the CO2 substitutes either species with a higher CCP or an organic solvent which, on completion of the process, may be partly recovered using energy-expensive technologies, or burned so as to

9

10

1 Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere Table 1.2

Some examples of technological applications of CO2.

Example of application

Product substituted

Example of application

Product substituted

Example of application

Product substituted

Dry cleaning

Chlorinated solvents

Fire extinguisher

Flame retardants

Extraction of caffeine

Hexane

Water treatment

Sulfuric acid

Cleaner in electronics

Fluorinated solvents

Solvent in reactions

Various organics

Air conditioning

Fluorinated compounds

Mechanical industry

Fluids with higher CCP

Production of nanomaterials

Organic solvents

Antibacterial

Complex pharmaceuticals

Extraction of fragrances

Hexane

Solvent in polymerization

Organic solvents

produce large amounts of CO2 [11]. Speciﬁc aspects of the use of CO2 as a technological ﬂuid are described in Chapter 2. Enhanced biological ﬁxation [10] has great potential, especially if an aquatic biomass can be created that is capable of producing chemicals and fuels; this topic is discussed in Chapter 13. The chemical utilization of CO2 is also considered in detail, with the pros and cons of each application highlighted to provide a detailed picture of the state of the art, and of the use of CO2. Areas in which further research in order to fully implement this utilization strategy are noted as key points. The potential of such a technology is especially interesting with regards to reducing CO2 emissions. In fact, often the most important aspect in the chemical utilization of CO2 is not the amount of CO2 used (if a chemical made from CO2 is used, the CO2 will be re-formed within a short time, except in the case of polymers that may last for decades!); rather, it is the introduction of innovative technologies that may lead to a reduction in the use of materials and energy. The potential uses of CO2 in chemical applications are shown in Figure 1.8, where some of the products (carboxylates, carbonates, and carbamates in routes A and B) are obtained by incorporation of the entire CO2 molecule. The reactions bearing such products will have a low energy content and may occur at room temperature, or lower (see Chapters 5 to 8). Processes in which CO2 is reduced to other C1 or Cn molecules (Figure 1.8, routes C and D) require an input of energy. In order to be consistent with the information provided in Section 1.5, such energy cannot be provided by fossil fuels, and alternative sources must be found, with solar energy the best candidate. In Chapters 9 to 12, details are provided of the potential for reducing CO2 to fuels or chemicals. Most processes involving CO2 conversion can be compared to the natural processes described in Chapter 3 (biological reactions) and Chapter 14 (geological inorganic reactions). The fundamentals of CO2 chemistry and reactivity are described in Chapter 4.

1.8 The Need for Research to Speed an Exploitation of the Utilization Option

11

O NH2

C

H2N

ONa/K

NH2

HO H3COH

COONa/K

A HCOOH

O

O

O

O

+

O O

O

O

O

n

O

D O

C

RO

HCONHR CnH2n+2

H2

CnH2n

CO CO 22

O ROH

RNH2

OR

CO COOH COOH

O

O2

H2C=CH2

O

B

O

HOOC

COOR

R

RC ≡ CR

Br

R

C

O

e-, H+

O

COOH

HOOC

O RNH2 + R’X

COOH

COOH

+

RNHC OR’ O

N H

Figure 1.8

O

C

H N N H

O

O OH

3

O

n

The possible applications of CO2 in chemical syntheses.

1.8 The Need for Research to Speed an Exploitation of the Utilization Option

The ﬁrst intensive investigations into the chemistry of CO2 followed the initial description of the transition-metal complex, Ni(CO2)(PCy3)2 [12]. Subsequently, during the late 1970s and 1980s, much emphasis was placed on investigating CO2 conversions and the associated reaction mechanisms [13, 14]. Unfortunately, however, this early enthusiasm in academia was not supported by industrial investment (at the time, there was no compulsion to reduce CO2 emissions and waste production!), and consequently during the following decades the research effort was decreased and the beneﬁts were lost. Yet, today, interest in CO2 utilization has been totally re-vamped, based on the above-mentioned aspects. Clearly, whilst current research must be funded along selected axes so as to identify the most

12

1 Carbon Dioxide: Utilization Options to Reduce its Accumulation in the Atmosphere

useful results in terms of exploitable technologies, it is vital that the fundamental research should also be continued. The use of CO2 in the synthesis of carboxylates (including acrylates), esters and lactones, carbonates, carbamates, and polymers may help to solve substantial problems encountered in the chemical industry, and hopefully advance towards sustainability. In all such cases, the new processes based on CO2 would replace old energy- and material-intensive technologies that employed either the use of toxic compounds (e.g., phosgene) or were characterized by a high E-factor [6]. The use of sc-CO2 as a solvent and reagent represents an application that reduce not only the amount of waste organic solvents, but also the emission of CO2 derived from the combustion of spent solvents. The dry-reforming of methane is a technology that may pressurize the gas-toliquid (GTL) approach by converting methane and CO2 into liquid fuels at the liqueﬁed natural gas (LNG) extraction well. The photoassisted electrochemical reduction of CO2 in water represents a very interesting technology that may allow the efﬁcient use of residual or intermittent energies [6], with the concomitant conversion of large volumes of CO2 into chemicals or fuels. Unfortunately, the technological use of CO2 has several hidden beneﬁts [6], the most prominent of which is that it should be implemented on a large scale so to replace toxic species and compounds which have a high CCP. The use of sc-CO2 has already been exploited in several areas (e.g., as a solvent in drywashing, or in extractions or chemical reactions); however, its use will inevitably be expanded into the chemical industry, as well as in the process of enhanced oil recovery. By mimicking Nature and combining biotechnology and chemistry, it may be possible to bring about the discovery of new technologies which, together with a greater use of biomass, may support the substitution of fossil-carbon with quasizero-emission technologies. One major exploitation of the concept that “Nature makes and Chemists re-shape” may bring about important beneﬁts that have not yet been fully identiﬁed. The “inorganication” of CO2 is a technology that may, at least potentially, be used to store large volumes of CO2 over the long term, in the form of “safe” chemicals. Of course, such an approach would be especially welcome in situations where residual inorganic oxides and sludge from industrial processes could be used for the CO2 ﬁxation. Whilst, in future, it is essential that industry, international and national funding organizations all support research investigations in the above-mentioned areas, it is equally vital that these studies are conducted in complementary fashion. In this way, it will be possible to investigate a wider variety of applications, taking into account speciﬁc factors applicable to different countries, and consequently producing the greatest beneﬁt to Society in general.

References

References 1 Mauna Loa Laboratory data, http://www. esrl.noaa.gov/gmd/ccgg/trends/. 2 Department of Energy (2008) International Energy Agency Report, 0484. 3 Solomon, S., Qin, D., Manning, M., Marquis, M., Averyt, K., Tignor, M.M.B., LeRoy Miller Jr, H., and Chen, Z. (2007) Climate Change 2007; The Fourth Assessment Report (AR4) of the United Nations Intergovernmental Panel on Climate Change (IPCC). 4 Aresta, M. (1990) Enzymatic and Model Carboxylation and Reduction Reactions for Carbon Dioxide Utilization (eds M. Aresta and J.V. Schloss), Kluwer Academic Publishers, Dordrecht, The Netherlands, p. 1. 5 Aresta, M., Quaranta, E., and Tommasi, I. (1991) Photochemical Conversion of Solar Energy (eds E. Pelizzetti and M. Schiavello), Kluwer Academic Publishers, Dordrecht, The Netherlands, p. 517. 6 Aresta, M. (2009) Carbon Dioxide Capture and Storage (ed. M. MarotoValer), Woodhead Publishing Limited, Abington Hall, Granta Park, Cambridge, CB21 6AH, UK. 7 The Conference Participants (1990) IEA Executive Conference on “Solar Photoconversion Processes for Recycling

8 9

10

11

12

13

14

Carbon Dioxide from the Atmosphere”, Colorado Spring. http://ceh.sric.sri.com/Public/Reports/ 743.2000. Graziano, P. (1990) Enzymatic and Model Carboxylation and Reduction Reactions for Carbon Dioxide Utilization (eds M. Aresta and J.V. Schloss), Kluwer Academic Publishers, Dordrecht, The Netherlands, p. 19. Aresta, M. (2010) Energy from aquatic biomass, in Handbook of Combustion (eds M. Lackner, F. Winter, and M. Agarwal), vol. 5, ch. 13, Wiley-VCH, Weinheim. Aresta, M., and Dibenedetto, A. (2002) CO2 Conversion and Utilisation (eds C. Song, A.M. Gaffney, and K. Fujimoto), ACS Series Book 809, American Chemical Society, p. 54. Aresta, M., Nobile, C.F., Albano, V.G., Forni, E., and Manassero, M. (1975) J. Chem. Soc. Chem. Commun., 15, 636. Aresta, M., and Forti, G. (eds) (1987) Carbon Dioxide as a Source of Carbon, Kluwer, Dordrecht, The Netherlands. Aresta, M., and Schloss, J.V. (1990) Enzymatic and Model Carboxylation and Reduction Reactions for Carbon Dioxide Utilization, Kluwer Academic Publishers, Dordrecht, The Netherlands.

13

15

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses Alessandro Galia and Giuseppe Filardo

2.1 Introduction

Studies of the use of liquid and supercritical carbon dioxide (scCO2) for chemical processes represent a vast ﬁeld which covers many interdisciplinary areas, including industrial chemistry, transport phenomena, catalysis, phase and chemical equilibria, and the design of high-pressure experimental and analytical apparatuses. In addition, CO2 can be used simply as a solvent or also as a reactant, thus further complicating the tassonomic approach to the problem. Another consideration that makes difﬁcult an original approach to analyzing the use of dense CO2 as a solvent for chemical synthesis is that overviews of the subject have been regularly provided in excellent reviews and books describing many different aspects of interest, including homogeneous and heterogeneous catalytic processes [1–12], polymerizations [13–23], green syntheses [24–27], and biocatalysis [28–31]. Consequently, the aim of this chapter is to highlight the main features of dense CO2 that permit the syntheses of both low-molecular-weight and macromolecular compounds. No attempt has been made to provide a comprehensive analysis of the literature in the different areas that, as noted previously, has been very effectively and regularly reviewed by others.

2.2 Dense Carbon Dioxide as Solvent Medium for Chemical Processes

A gaseous pure component can be deﬁned as supercritical when its state is determined by values of temperature T and pressure P that are above its critical parameters (Tc and Pc). In the proximity of its critical point, a pure supercritical ﬂuid (or a dense gas as it is alternatively known) has a very high isothermal compressibility, and this makes possible to change signiﬁcantly the density of the ﬂuid with relatively limited modiﬁcations of T and P. On the other hand, it has been shown that the thermodynamic and transport properties of supercritical ﬂuids can be tuned simply by changing the density of the medium. This is particularly interesting for Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

16

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses

the chemical engineer, as density becomes an additional parameter that can be used to improve the performances of a process. Supercritical carbon dioxide was initially mainly investigated as a processing medium for extractions and fractionations in the ﬁeld of natural product processing. Subsequently, the past ﬁfteen years has witnessed a rapidly growing interest in the use of this unconventional solvent medium to conduct chemical reactions, mainly because it can serve as a substitute for liquid organic solvents [32]. Such interest in supercritical ﬂuids has been shown to depend not only on their attractive chemical– physical properties, such as density-adjustable solvent strength, low surface tension and viscosity, but also on speciﬁc technical–economical properties, such as low cost, high availability, high stability, good biocompatibility, low toxicity and nonﬂammability, and readily accessible critical parameters (Tc = 304.12 K, Pc = 7.37 MPa). In particular, when gaseous reagents are involved, scCO2 can allow the operator to perform the reaction under single-phase conditions, thus greatly increasing the local concentration of the dissolved gases and markedly accelerating the mass transfer kinetics both for the elimination of the gas–liquid interphase and for the better transport properties of the supercritical medium in comparison with liquid solvents. These considerations may explain why the most extensively investigated reactions using scCO2 are hydroformylation, hydrogenation, and oxidation. The hydroformylation reaction (Scheme 2.1), originally discovered by Otto Roelen in 1938 [33], is an industrial process of strategic importance for the manufacture of aldehydes from oleﬁns and syngas. Most research on the hydroformylation of oleﬁns in scCO2 has been carried out using homogeneous catalytic systems. From an applications point of view, it is interesting to note that good results have been obtained with high-molecularweight oleﬁns which, owing to their poor water-solubility, cannot be hydroformylated in the aqueous liquid/liquid/gas system based on the use of a water-soluble rhodium (Rh) catalyst (Ruhrchemie/Rhône-Poulenc). Rh has frequently been adopted as the metal center in scCO2 and one crucial point that must be resolved for transferring the synthetic route from the laboratory to the industrial plant, is the availability of low-cost CO2-philic ligands to render the organometallic catalyst soluble in the reaction medium at a sufﬁciently high concentration. This problem is particularly difﬁcult to overcome because the solvent power of CO2 will become progressively reduced, as the result of antisolvent effects, when the molar fractions of low-critical-temperature gaseous reagents are increased [34–36]. To date, the most effective approach to this problem has been to functionalize the aryl ring of the triphenylphosphine with ﬂuorinated alkyl groups [37–41];

COH R Scheme 2.1

+ CO+H2

catalyst

COH R

+

R

Schematic representation of the hydroformylation of terminal alkenes.

2.2 Dense Carbon Dioxide as Solvent Medium for Chemical Processes

however, this solution proved to be too expensive for industrial applications unless a quantitative recovery of the catalytic system could be achieved. More recently, Sarbu et al. showed that the incorporation of carbonyl groups could signiﬁcantly increase the solubility of macromolecular materials in scCO2, provided that the enthalpic and entropic contributions were properly balanced [42, 43]. This result was subsequently extended to the design of carbonylated phosphines that were tested as ligands in the Rh-catalyzed hydroformylation of 1-decene in scCO2 [44]. Phosphite, phosphonite and phosphinite ligands with nine-carbon branched alkyl substituents have also been synthesized, and proved to be active in the rhodiumcatalyzed hydroformylation of 1-octene in scCO2, although the catalytic system was insoluble in the dense gas [45]. Heterogeneously catalyzed processes have also been the object of much investigation. In this case the problem of catalyst recycling was considerably reduced and the excellent transport properties of the supercritical medium still granted rapid mass transfer rates. As an example, interesting results were obtained by the group of Poliakoff, when they studied the hydroformylation of 1-octene using a Rh complex with aryl phosphine ligands immobilized on silica [46]. The reasons for studying hydrogenation reactions in scCO2 are similar to those described previously for oxo-syntheses, where the complete miscibility of molecular hydrogen with scCO2 was invoked as an important precondition to achieve process intensiﬁcation. In this context, it should be noted that the hydrogenation of organic substrates using continuous ﬁxed-bed reactors is one of the few processes that have been developed effectively from the laboratory scale to the industrial application. This was achieved on the basis of an excellent collaboration between the Clean Technology Group at the University of Nottingham and the ﬁne chemicals manufacturer Thomas Swan & Co. Ltd [47]. Despite the above-mentioned information relating to gaseous reactant solubility seeming to push towards achieving a one-phase reaction system, it has been shown [48] that faster reaction rates can occasionally be obtained under biphasic conditions. Notably, this occurs when CO2 dissolved in the organic reactants permits a signiﬁcant increase in the solubility of permanent gases in the expanded liquid phase. This, in turn, allows the problems associated with low reactant concentrations, which are often observed in single-phase processes and which nullify the high local concentration of hydrogen, to be overcome. It is important to note that the formation of CO from the reverse water gas shift reaction (CO2 + H2 = CO + H2O) has been detected at low temperature with the commonly used Pt group metals [49, 50]. Under the operational conditions used to carry out these reactions, the CO coverage is relatively low and concentrated on speciﬁc sites such as the step and kink. This has been considered a possible explanation for the fact the catalytic performances are generally unaffected unless demanding reactions that require the catalytic action of speciﬁc sites on the catalyst surface are considered, as is the case with enantioselective hydrogenation reactions [49]. Another important class of reactions that has been the object of intense investigation is the selective oxidation of organic substrates in dense CO2. Clearly, the ﬁrst consideration supporting the use of CO2 as solvent for such reactions is that,

17

18

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses HO

O2 catalyst Scheme 2.2

O

+

+ H2O

Oxidation of cyclohexane to a mixture cyclohexanol/cyclohexanone.

in contrast to most adopted organic solvents, this compound cannot be further oxidized, thus preventing the formation of byproducts and waste materials. Many of the above considerations relating to hydroformylation and hydrogenation reactions in CO2 can be extended to oxidations, particularly when O2 is used as the oxidant agent. Indeed, O2 is the most widely investigated oxidizing compound, due to its low cost and high atom-efﬁciency. In these processes, signiﬁcant process advantages have been observed arising from the reduction of transport resistance through elimination of the gas–liquid interface. This is usually coupled to a higher local concentration of the oxidizing agent, accompanied by a better safety proﬁle of the process due to the solvent acting as inert diluent. In this context, it is of interest to note dense CO2 is one of the few reaction media that can be used for the direct synthesis of hydrogen peroxide from H2 and O2 [51]. The majority of studies on oxidation reactions in scCO2 have involved catalyzed processes promoted by molecular oxygen, in which the role of the catalyst is to generate free radicals that will react with the chemical oxidant, leading to a product distribution that is typical of an unselective chain process. Among these can be mentioned the oxidation of cyclohexane to cyclohexanol and cyclohexanone (Scheme 2.2) as an intermediate step in the production of the adipic acid that is a key component in the production of Nylon 6,6 polyamide [52–54]. Oleﬁn epoxidation is another synthetic route of great interest that has been the object of several investigations, not only in multiphase reaction systems but also with oxidants other than molecular oxygen, such as inorganic and organic peroxides [55–57]. The partial oxidation of alcohols, to afford carbonyl or carboxylic compounds, is another synthetic route of high industrial interest. For this, scCO2 was investigated as a reaction medium for the aerobic oxidation of aliphatic, unsaturated, aromatic and benzylic acids with different catalytic systems, mainly based on the use of noble metals, both in batch [58–64] and in continuous ﬁxed-bed reactors [65–70]. In this context, very promising results have been obtained when studying the catalytic activity of supported palladium and gold nanoparticles in the oxidation of benzyl alcohol to benzaldehyde; these allowed conversions and selectivities in excess of 90% to be achieved [71–73].

2.3 Enzymatic Catalysis in Dense Carbon Dioxide

Enzymes are natural biomacromolecules that belong to the class of proteins. In an aqueous environment, they assume conformations that impart them with high

2.4 Other Reactions in Dense Carbon Dioxide

activities and selectivities under mild conditions. Although, in the natural world, enzymes have adapted to function in aqueous biological media, they have also been shown capable of functioning in organic solvents, provided that a small number of water molecules are bound to the biocatalyst; this allows the enzyme to exist in its active conformation [74]. As enzymes are insoluble in organic media, they represent a special class of heterogeneous catalysts. The use of enzymatic catalysts in scCO2 would seem the perfect union between a highly selective and active green catalytic system and an environmental friendly solvent with excellent transport properties. Such a union could achieve a marked reduction in the mass transfer resistance that is, most likely, the slowest step in a heterogeneously catalyzed process carried out with a chemically active catalytic system. For these reasons, many groups have investigated CO2 enzymatic catalysts that have been either freely suspended or stabilized by using different approaches [31]. Interestingly, the hydrophobic ion pairing of cytochrome c with ﬂuorinated anionic surfactant proved to be an effective strategy for preparing enzyme–surfactant complexes that were soluble in dense CO2, without any modiﬁcation of the secondary structure of the protein [75]. Despite, in theory, this approach having a high potential for this reaction, the practical use of biocatalysts in scCO2 was frequently – but not always – prevented by solvent–catalyst interactions that led to the formation of carbamates. Such a reaction changes the enzyme’s conformation, and in turn affect its catalytic performances. As a result, the industrial use of these catalysts presents several problems, the solutions of which are partially dependent on the ability to identify enzymes that could be obtained from organisms adapted to living in a CO2-rich environment.

2.4 Other Reactions in Dense Carbon Dioxide

Whilst a major proportion of studies have been targeted at using scCO2 as a green solvent in organic syntheses, and have focused on hydrogenation, hydroformylation and oxidation reactions, other classes of organic reactions have been studied in dense carbon dioxide. Some of these were mainly used to assess density inhomogeneity effects that, in the proximity of the critical point, can markedly affect the kinetics of reactions carried out in a supercritical ﬂuid (SCF), as in the case of Diels–Alder reactions [76–83]. Other reactions were investigated within the frame of cooperation between academia and industry, and led to patents that were the object of industrial development, as was the case of acid-catalyzed Friedel–Craft reactions [84, 85]. Apart these cases, many other reactions have been investigated, including carbonylations, Heck reactions, vinyl substitutions, and hydrosylation. In general, based on available data, it is difﬁcult to establish whether these studies could extend beyond the laboratory scale. However, as Eric Beckman has stressed, when reviewing the use of supercritical or near-critical carbon dioxide in “green” chemical syntheses [26], the role of these studies in demonstrating the extreme versatility of

19

20

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses

dense CO2 as a solvent for chemical syntheses, thus stimulating new approaches that could become relevant for industrial applications, cannot be neglected.

2.5 Polymer Synthesis in Supercritical Carbon Dioxide

The primary consideration is that polymer synthesis in SCFs are generally carried out at high pressure; consequently, it is important to analyze the inﬂuence of this parameter on the polymerization process. The increase of pressure can affect a polymerization process by:

• • •

increasing the concentration of gaseous monomers; changing the kinetic constants for the individual steps involved in the process; modifying the equilibrium constant of the process.

Apart from pressure effects, which are not unique to SCFs, some other peculiar aspects of polymerization in SCFs are related to the physico-chemical properties of compounds under supercritical conditions. In traditional liquid solvents, the polymerization reaction rates are often limited by the local increase in viscosity during the process, as this lowers the mass transfer rate of the monomer to the reaction site. A lower viscosity and a higher diffusion coefﬁcient in SCFs each contribute to overcome this limitation, however, allowing the polymerization rate to be signiﬁcant up to high value of monomer conversion. The initiation step could also be positively affected by the above-mentioned transport properties, as the efﬁciency factor f assumes higher values with respect to conventional liquid solvents due to the diminished solvent cage effect. One further advantage is constituted by the tunability of the compressibilitydependent properties such as density, dielectric constant, heat capacity, and viscosity, all of which offer additional possibilities to modify the performances of the polymerization process. This aspect could be particularly relevant in the case of copolymerization reactions, where the reactivity ratios of the two monomers, and ultimately the ﬁnal composition of the copolymer, could be controlled by modifying the pressure of the reaction system. Apart from polymerization processes with gaseous monomers above their critical points – for example, the synthesis of low-density poly(ethylene) – several SCFs have been tested as inert reaction media, such as ethane, propane, butane, and CO2. Among these, scCO2 is by far the most widely investigated, because it links positive ﬂuid effects on the polymers with environmental advantages; this makes scCO2 the main candidate as an alternative to traditional solvents used in polymer syntheses. Compared to water – another green solvent which is frequently adopted as a dispersing medium in industrial heterogeneous polymerizations – scCO2 eliminates the need for energy-intensive drying processes, as the polymer product can

2.5 Polymer Synthesis in Supercritical Carbon Dioxide

be isolated completely dry upon venting. As the critical temperature of CO2 is very close to room temperature, this solvent can be used in applications that involve polymers and heat-sensitive materials such as enzymes, ﬂavors, pharmaceuticals, and highly reactive monomers. With regards to the physico-chemical properties, the solvency of scCO2 as a medium and its plasticizing effects on macromolecular materials are of central importance. The solubility of reagents and polymer product in the continuous phase of a polymerization medium is the ﬁrst parameter to consider in the selection of the polymerization technique (whether homogeneous or heterogeneous). CO2 has a low dielectric constant, and by varying T and P, its value can be changed from 1.01 to 1.45 for the gas phase and to 1.60–1.67 for the liquid phase. It has been noted that CO2 behaves very much like a hydrocarbon solvent with regards to its ability to dissolve small molecules; consequently, many monomers exhibit a high solubility within CO2. On the other hand, most high-molar-mass polymers are scarcely soluble in CO2, and the only polymers that show good solubility under relatively mild conditions (T < 373 K, P < 35 MPa) are amorphous ﬂuoropolymers, silicones, and polyether polycarbonate copolymers. To date, the parameters that govern the solubility of polymers in CO2 are not fully understood, despite several investigations having been conducted to clarify the nature of speciﬁc solute–solvent interactions between various polymers and CO2. In the case of silicones, a speciﬁc interaction between CO2 and the silicone portion of the polymer backbone determines the high solubility of these types of macromolecular compound. Lewis acid–base-type interactions have been detected by using Fourier transform infrared (FTIR) spectroscopy between CO2 and the electron-donating functional groups of polymer chains [e.g., the carbonyl group of poly(methyl methacrylate)] and high-pressure 19F nuclear magnetic resonance (NMR) studies have indicated speciﬁc interactions between the CO2 and ﬂuorocarbons. Quite recently, Sarbu et al. [42, 43] proposed a notable thermodynamic approach for the rational design of macromolecules with high solubility in scCO2. The basic concept of such a soluble macromolecule was to select its constituent repeat units so as to induce favorable enthalpic and entropic interactions with the compressible ﬂuid to ensure a high gain in the free energy change of mixing: ΔGmix = ΔH mix − TΔSmix . This conceptual design was successfully used as a guide to identify nonﬂuorous, nonsilicon polymers that were soluble in CO2 [86, 87]. Regardless of the exact nature of the polymer–solvent interactions, it is clear that the poor solubility of most polymers in CO2 compels their syntheses to be performed via heterogeneous polymerization techniques to operate at acceptable values of pressure and temperature. Owing to the good solubility of many monomers and initiators in CO2, dispersion polymerization is today the most investigated heterogeneous method used for the synthesis of high-molecular-weight CO2-insoluble polymers. One central role for the performance of a successful dispersion polymerization is the surfactant.

21

22

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses

Since traditional surfactants (which have been designed for use in aqueous or organic continuous phases) are scarcely soluble in CO2, specialized stabilizer compounds have been developed that are capable of stabilizing the polymer colloid. All of these molecules are characterized by a CO2-philic portion which shows high solubility in the solvent, and a CO2-phobic portion which has a high afﬁnity to the polymer phase. The latter part, which constitutes the anchoring portion of the surfactant, attaches to the surface of the polymer particle by either physical adsorption or chemical grafting, depending on its nature. The CO2-philic portion extends into the continuous phase and prevents the onset of ﬂocculation of the particles by imparting long-range repulsion among them. However, these effects must be strong enough to compensate the long-range van der Waals attractions. Beside the consideration of solubility, another property of scCO2 which is relevant in polymer synthesis is its capability to highly plasticize polymers, thereby lowering their glass transition temperature (Tg). This plasticization effect is especially important in heterogeneous processes where a polymer coagulum with a high local viscosity is formed, as this effect may lead to an increase in the free volume of the polymer chains that induce an enhancement in the value of the diffusion coefﬁcient of species inside the coagulated particles. Under these circumstances, the polymerization process can proceed up to high value of monomer conversion. The translational and segmental diffusivity of the growing polymer chains are also enhanced by the plasticizing effect, and this may lead to a shift in the onset of the gel effect at a higher value of the conversion or, occasionally, to a practical suppression of such an effect. Nonetheless, such behavior can simplify the thermal control of the process. 2.5.1 Chain Polymerizations: Synthesis of Fluoropolymers

Several monomers have been polymerized in scCO2 by particle-forming chain free-radical polymerizations, the most investigated of these being methylmethacrylate, styrene, N-vinylpyrrolidone, acrylic acid, vinylidene ﬂuoride, and tetraﬂuoroethylene [13–23]. Among these monomers, those of major interest in terms of the industrial use of the synthetic route are ﬂuorinated vinyl monomers, as these allow the preparation of high-performance polymers that exhibit a unique combination of excellent chemical resistance, a high thermal stability, a low dielectric constant and dissipation factor, unusual surface properties, low water absorptivities, excellent weatherability, and low ﬂammabilities. These positive properties are related to the presence of ﬂuorine substituents in the chain, although this may lead to a more difﬁcult control of the polymerization process when compared to their hydrocarbon homologues. In fact, the high electrophilicity of the ﬂuorinated macroradical species that propagate the chain represents the origin of a strong kinetic competition between chain propagation and chain transfer reactions involving H-atom donor species [88]. Fluoropolymers are usually obtained through heterogeneous (emulsion or suspension) polymerization techniques in aqueous systems. As a consequence of the

2.5 Polymer Synthesis in Supercritical Carbon Dioxide

adopted aqueous initiators, thermally unstable end groups (carboxylic acid and acid ﬂuoride) are generated that must be removed before the polymer melt can be processed. In order to reduce the amount of thermally unstable end groups, the polymerization could be performed in aprotic solvents such as chloroﬂuorocarbons (CFCs) or perﬂuorocarbons, perﬂuoroalkyl sulﬁde and perﬂuorinated cyclic amines; however, these solvents are not only prohibitively expensive but also nonenvironment friendly. In this context, dense CO2 represents a sustainable alternative to such volatile organic compounds (VOCs), as demonstrated implicitly by the huge investments made over the past few years, both in the USA (i.e., the cooperation of DuPont with University of North Carolina) and in the European Union (European Research Grants Superpol and Ecopol), in investigating the use of scCO2 for the production of ﬂuorinated polymers. High-molecular-mass polytetraﬂuoroethylene (PTFE) was synthesized in a heterogeneous CO2/aqueous medium at 5–12 MPa, using a water-soluble persulfate initiator, with and without a sodium or ammonium salt of perﬂuoro-octanoic acid (C7F15COO−Na + NH+4 ) as surfactant (to stabilize the water-rich micelles) [89]. In this process, by correctly selecting a water-soluble perﬂuorinated initiator, the compartmentalization of monomer, polymer and initiator was achieved, thus obtaining rapid polymerization kinetics and high values of both yields (up to 90%) and molecular weights (Mn = 900 kg mol−1) [89]. Moreover, the monomer/CO2 system behaved like a pseudoazeotropic mixture and could be safely handled. In fact, as TFE is known to form explosive mixtures with air, it must be kept scrupulously free from oxygen so as to prevent the formation of any shock-sensitive polymeric peroxides. Moreover, in the absence of air, it can disproportionate violently to form CF4 and elemental carbon [90]: (CF2=CF2 → C + CF4). More recently, the polymerization of TFE was carried out in a “water-free” scCO2-based medium, at 323 K and 12.1–13.3 MPa, with tertiary-amyl-per-pivalate as a free radical initiator, using a 5 liter batch reactor [91]. The experiments were performed both in the absence and presence of perﬂuoropolyether carboxylate, employed as surfactant, under the free acid form and as sodium and calcium salts (Figure 2.1). An expanded ﬁbrillated PTFE morphology, which probably arose from the plasticizing effect of scCO2 on the amorphous domains of the polymer, was obtained in all experiments. Such a morphology could be particularly interesting for applications in the manufacture of solvent-free hydrophobic microporous membranes. In the case of surfactant-free polymerizations, the product was mainly constituted by irregular macroparticles having sizes in the range of 200 to 500 μm. Interestingly, when the free acid was used a clear acceleration of the polymerization rate was observed, and the polymer was obtained in the form of smaller particles with a more regular quasi-spherical shape. Poly(vinylidene ﬂuoride) (PVDF) is the second most important thermoplastic within the ﬂuoropolymer family after PTFE. Although, both the thermal and chemical stability of PVDF are somewhat lower compared to PTFE, the hydrogenated polymer can be more easily processed with conventional equipment, and it offers an advantageous compromise between quality and price. When the

23

24

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses

FLUOROLINK® C stabilizers CF3 CF2COO-NH4+

Cl CF2CFO

7004A n=4-6, Mw=937

n CF3 CF2COO- Ca2+

Cl CF2CFO

n

7004Ca n=4-6, Mw=1878

2

CF3 Cl CF2CFO

CF2COOH

7004H n=4-6, Mw=920

n Figure 2.1 Perﬂuoropolyether (PFPE) carboxylate compounds used as stabilizers in the polymerization of tetraﬂuoroethylene in scCO2.

synthesis of PVDF in scCO2, using a continuous stirred tank reactor, was recently investigated [92–95] the synthesized polymer showed properties similar to those of commercial samples prepared by the conventional route. The precipitation polymerization of VDF in dense carbon dioxide, initiated by γ-radiation, was studied as part of the aforementioned Superpol EU research program [96]. The absence of any chemical initiators, the fragments of which would ultimately be incorporated into the polymer chains, together with the easy removal of volatile carbon dioxide and monomer, resulted in a high-purity product obtained under mild process conditions (T ≤ 313 K, P ≤ 25 MPa). A further improvement in the process could be achieved performing the synthesis of ﬂuorinated monomers in scCO2 in the presence of suitable stabilizers, so as to modify the polymerization technique from precipitation to dispersion or emulsion. When a research group at the University of Nottingham investigated the polymerization of VDF in the presence of a reactive polydimethylsiloxane monomethacrylate (PDMS-mMA) [97], even when an increased polymer molecular weight was obtained, a high level of aggregation of the primary particles was observed. The same group achieved a more effective stabilization by using a home-synthesized ﬂuorinated graft maleic anhydride copolymer stabilizer [98]. In this case, the formation of spherical microparticles was clearly observed, but a monomer conversion less than 15% was reported, which corresponded to a maximum solid content of polymer in the autoclave of approximately 90 g l−1. Other groups [99, 100] investigated the possibility of carrying out a surfactantassisted polymerization of VDF in scCO2, using diethyl peroxydicarbonate (DEPDC) as a free radical initiator, with the aim of testing several perﬂuoropolyether compounds synthesized by Solvay Solexis as stabilizers. When ammonium carboxylate derivatives were used, an effective stabilization of the polymer particles was obtained, and the polymer was collected from the reactor with yields of up to 63%. This corresponded to a solid content in the reactor of 220 g l−1, in the form

2.5 Polymer Synthesis in Supercritical Carbon Dioxide

of a powder constituted by spherical particles, the diameters of which decreased as a function of the concentration of the stabilizer. Both, TFE and VDF may be copolymerized with suitable comonomers in order to obtain macromolecular materials with improved end-use properties. One ﬂuorinated monomer that was copolymerized under mild conditions (308 K and 9–11 MPa) was perﬂuoropropyl-vinylether (PPVE) (CF3CF2CF2OCF=CF2). The copolymer was obtained in the form of a free-ﬂowing powder, with almost quantitative yields; moreover, its composition could be altered by changing the molar ratio of the comonomers in the feed. Interestingly, chains were synthesized with molecular weights greater than that of the commercial product with a concentration of carboxylic acid or acid ﬂuoride end groups similar to those detected in polymers prepared in CFCs [101]. Macromolecular matrixes suitable for the preparation of perﬂuorinated ionexchange membranes can be obtained through the copolymerization of TFE and functionalized perﬂuoro-vinyl-ether monomers, such as sulfonyl-vinyl ﬂuoride CF2=CFOCF2CF(CF3)O(CF2)2SO2F (PSEVPE). These copolymers should couple to provide a good mechanical strength with a high ionic conductivity. In order to achieve this goal, the polymer must have a high molecular weight and a high concentration of ionic groups grafted onto the backbone. The main drawback in the synthesis of such a polymer with these joined properties is constituted by the β-scission of vinyl ether-terminated macroradicals [101]. Poly(TFE-co-PSEVPE) can be synthesized in scCO2 with a concentration of ionic groups similar to those obtained in conventional routes, but virtually free of acid end groups as a result of a substantial suppression of the β-scission reactions. Copolymers of TFE and 2,2-bis(triﬂuoromethyl)-4,5-diﬂuoro-1,3-dioxole (PDD) were synthesized in CO2 at low temperatures using bis(perﬂuoro-2-N-propoxypropionyl) peroxide as initiator [102, 103]. In this case, the copolymers were prepared with different compositions and having a broad range of Tg-values, from 340 to 607 K. No difference could be found between a ﬂuorinated commercial sample (Teﬂon AF 1601) and a copolymer synthesized in CO2 having a similar composition, by using 19F NMR spectroscopy, IR spectroscopy, and differential scanning calorimetry (DSC). TFE may be copolymerized in dense CO2 also with hydrocarbon vinyl monomers such as vinyl acetate (VAc). In this case, it is possible to prepare an organic solventsoluble ﬂuoropolymer that might be functionalized in a further step, for example through the hydrolysis of VAc groups to vinyl alcohol groups [104, 105]. Poly(TFEco-VAc) can be synthesized in an aqueous medium (e.g., by emulsion), although a branched structure is obtained via this route due to hydrogen abstraction from the VAc groups. In contrast, copolymers produced in scCO2 (318 K, 23–26 MPa initial pressure) are predominantly linear, most likely due to the swelling properties of CO2 towards ﬂuoropolymer matrices; this facilitates the diffusion of monomers into the polymer phase, thus increasing the rate of propagation (bimolecular process) relative to the intramolecular chain transfer (unimolecular process). Among VDF copolymers, those that include hexaﬂuoropropylene (HFP) have an important commercial value. The properties of the copolymer depend on the

25

26

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses

HFP content; when it is less than 19–20 mol% a thermoplastic and semicrystalline macromolecule is obtained, whereas for higher HFP contents the copolymers are amorphous and elastomeric. Most commercial poly(VDF-co-HFP) elastomers contain approximately 22 mol% HFP, a composition that represents the best compromise between a low Tg and a fully amorphous polymer. These materials may also be produced commercially via suspension and emulsion polymerization in water. Both processes generate large quantities of wastewater, and have high energy costs for drying of the polymer. Moreover, ammonium perﬂuoro-octanoate surfactants, which are environmentally suspect, are frequently adopted. In the case of copolymers, consideration must be taken of the effect of chain composition on the solubility of the macromolecules in scCO2. For poly(VDF-coHFP), it has been proposed that the copolymer is insoluble in the polymerization medium when the HFP content is low, but it becomes soluble at a high HFP content [106]. Interestingly, the composition of the copolymer can be controlled by altering the monomer feed composition. For example, when VDF and HFP are copolymerized, two types of propagating macroradicals are possible – one with VDF at the propagating end, and the other with HFP. The reactivity ratios were estimated by performing batch polymerizations in scCO2 at low values of monomer conversion, and analyzing the copolymer composition with 19F NMR. The HFP was found to undergo cross-propagation with VDF, while the latter self-propagated randomly and also cross-propagated with HFP [107, 108]. It was also noted that the copolymer was always less rich in HFP with respect to the feed composition. From an industrial viewpoint, it is interesting to note that poly(VDF-co-HFP) can be synthesized via a surfactant-free copolymerization in scCO2, and would have properties similar to those of the copolymers obtained via an aqueous emulsion copolymerization. Advantageously, however, this approach would avoid wastewater streams, reduce the energy costs (no drying step is required), and also improve the chemical stability of the polymers due to the absence of reactive acid ﬂuoride end groups. 2.5.2 Step Polymerizations: Synthesis of Biodegradable Polymers

Although the main body research into polymerization in scCO2 has been focused on chain free radical polymerization processes, step growth polymerizations have also been object of much investigation. In these processes, low-molecular-weight molecules are often coproduced that must be removed in order to shift the equilibrium towards the synthesis of high-molar-mass polymers. However, the increase in viscosity that accompanies the growth of the chains also lowers the diffusivity of such small molecules, which makes the process mass transfer-controlled. But, scCO2 can be used to overcome this drawback since, by plasticizing the polymer phase, it induces an acceleration of the kinetics of diffusion and of the mobility of the polymer chain in the melts. Within the frame of these processes, it is interesting to note that scCO2 appears to serve as an effective dispersion medium for the synthesis of biodegradable poly-

Acknowledgments

mers such as poly(D,L-lactide-co-glycolide) and poly(ε-caprolactone). These macromolecular materials can be used in biomedical applications, an example being the preparation of a “scaffold” for tissue growth. However, given their end-use, it is crucial that these materials are free from any residues of the catalyst, monomer, and solvent. In this regard, scCO2 may represent a valuable solution, as it can be used not only as a nontoxic dispersion medium for the heterogeneous ringopening polymerization (ROP) of lactide, glycolide and ε-caprolactone, but also as an extraction solvent to recover any unreacted monomers. Some interesting results have been reported for these polymerization processes, both in the absence [109, 110] and presence of a suitable surfactant [111–114]. Apart from potentially toxic Lewis acid catalysts, enzymes have also been used successfully to perform the ROP of ε-caprolactone [115].

2.6 Conclusions

The beneﬁts that scCO2 offers as a reaction medium have been the object of intense research activity, and it is foreseeable that this ﬁeld of study will continue to improve. This suggestion is justiﬁed by the fact that scCO2 demonstrates both favorable physico-chemical properties (tunability of the chemico-physical properties, complete miscibility with permanent gases, plasticizing effect of polymers, inertness to free chain reactions) and positive technical and environmental qualities (naturally occurring and abundant, inexpensive, nontoxic, nonﬂammable). In many cases, scCO2 may offer not only an easier separation of the product but also an improved safety proﬁle for a process; examples of this are the oxidation reactions or polymerization of TFE. At present, the majority of these research studies are still conﬁned to the laboratory scale, with transfers to the industrial scale – such as the DuPont process for the synthesis of ﬂuorinated polymers in scCO2, or the Thomas Swan plant for continuous reactions in dense CO2 – still quite rare. Whilst this diversity is partly due to the natural difﬁdence that envelops a new technology, it is also due to the fact that, in many cases, the environmental beneﬁts imparted by the use of CO2 are not accompanied by clear technical– economical advantages, which are the main driving forces of industry. Yet, it is clear that a more integrated interdisciplinary cooperation among chemists, engineers, and biologists could play an important role in the development and promotion of more promising processes utilizing scCO2.

Acknowledgments

Part of the research reported in this chapter was supported by the Ministero dell’Istruzione, dell’Università e della Ricerca (MIUR) and by the Università di Palermo.

27

28

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses

References 1 Subramaniam, B. and McHugh, M.A. (1986) Ind. Eng. Chem. Process Des. Dev., 25, 1–12. 2 Kaupp, G. (1994) Angew. Chem., Int. Ed. Engl., 33, 1452–1455. 3 Baiker, A. (1999) Chem. Rev., 99, 453–473. 4 Kruse, A. and Vogel, H. (2008) Chem. Eng. Technol., 31, 1391–1395. 5 Jessop, P.G., Ikariya, T., and Noyori, R. (1995) Science, 269, 1065–1069. 6 Jessop, P.G., Ikariya, T., and Noyori, R. (1999) Chem. Rev., 99, 475–493. 7 Jessop, P.G. and Leitner W. (eds) (1999) Chemical Synthesis Using Supercritical Fluids, Wiley-VCH Verlag GmbH, Weinheim. 8 Jiang, H. (2005) Curr. Org. Chem., 9, 289–297. 9 Jessop, P.G. (2006) J. Supercrit. Fluids, 38, 211–231. 10 Nunes da Ponte, M. (2009) J. Supercrit. Fluids, 47, 344–350. 11 Leitner, W. (2000) Appl. Organomet. Chem., 14, 809–814. 12 Campestrini, S. and Tonellato, U. (2005) Curr. Org. Chem., 9, 31–47. 13 Scholsky, K.M. (1993) J. Supercrit. Fluids, 6, 103–127. 14 Kiran, E. (1994) Supercritical Fluids (eds E. Kiran and J.M.H. Levelt Sengers), Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 541–588. 15 DeSimone, J.M., Maury, E.E., Menceloglu, Y.Z., McLain, J.B., Romack, T.J., and Combes, J.R. (1994) Science, 265, 356–359. 16 Shaffer, K.A. and DeSimone, J.M. (1995) Trends Polym. Sci., 3, 146–153. 17 Cooper, A.I. and DeSimone, J.M. (1996) Curr. Opin. Solid State Mater. Sci., 1, 761–768. 18 Canelas, D.A. and DeSimone, J.M. (1997) Adv. Polym. Sci., 133, 103– 140. 19 Kendall, J., Canelas, D.A., Young, J.L., and DeSimone, J.M. (1999) Chem. Rev., 99, 543–563. 20 Cooper, A.I. (2000) J. Mater. Chem., 10, 207–234.

21 Ajzenberg, N., Trabelsi, F., and Recasens, F. (2000) Chem. Eng. Technol., 23, 829–839. 22 Wood, C.D., Cooper, A.I., and DeSimone, J.M. (2004) Curr. Opin. Solid State Mater. Sci., 8, 325–331. 23 Du, L., Kelly, J.Y., Roberts, G.W., and DeSimone, J.M. (2009) J. Supercrit. Fluids, 47, 447–457. 24 Subramaniam, B., Lyon, C.J., and Arunajatesan, V. (2002) Appl. Catal. B Environ., 37, 279–292. 25 Leitner, W. (2002) Acc. Chem. Res., 35, 746–756. 26 Beckman, E.J. (2004) J. Supercrit. Fluids, 28, 121–191. 27 Liu, S. and Xiao, J. (2007) J. Mol. Catal. A Chem., 270, 1–43. 28 Nakamura, K. (1990) Trends Biotechnol., 8, 288–292. 29 Kamat, S.V., Beckman, E.J., and Russel, A.J. (1995) Crit. Rev. Biotech., 15, 41–71. 30 Matsuda, T., Harada, T., and Nakamura, K. (2005) Curr. Org. Chem., 9, 299–315. 31 Hobbs, H.R. and Thomas, N.R. (2007) Chem. Rev., 107, 2786–2820. 32 Niessen, H.G. and Woelk, K. (2007) Top. Curr. Chem., 276, 69–110. 33 Roelen, O. (1938) German Patent 849 548. 34 King, J.W. and Zhang, Z. (1998) Anal. Chem., 70, 1431–1436. 35 Roth, M. (1998) Anal. Chem., 70, 2104–2109. 36 Cipollina, A., Anselmo, R., Scialdone, O., Filardo, G., and Galia, A. (2007) J. Chem. Eng. Data, 52, 2291–2297. 37 Palo, D.R. and Erkey, C. (1998) Ind. Eng. Chem. Res., 37, 4203–4206. 38 Koch, D. and Leitner, W. (1998) J. Am. Chem. Soc., 120, 13398–13404. 39 Palo, D.R. and Erkey, C. (1999) Ind. Eng. Chem. Res., 38, 786–3792. 40 Banet Osuna, A.M., Chen, W., Hope, E.G., Kemmitt, R.D.W., Paige, D.R., Stuart, A.M., Xiao, J., and Xu, L. (2000) J. Chem. Soc. Dalton Trans., 4052–4055. 41 Koeken, A.C.J., van Vliet, M.C.A., van den Broeke, L.J.P., Deelman, B.J., and

References

42 43

44 45

46

47

48

49

50

51 52

53

54

55

56

57

Keurentjes, J.T.F. (2006) Adv. Synth. Catal., 348, 1553–1559. Sarbu, T., Styranec, T., and Beckman, E.J. (2000) Nature, 405, 165–168. Sarbu, T., Styranec, T., and Beckman, E.J. (2000) Ind. Eng. Chem. Res., 39, 4678–4683. Hu, Y., Chen, W., Xu, L., and Xiao, J. (2001) Organometallics, 20, 3206–3208. Giménez-Pedrçs, M., Aghmiz, A., Ruiz, N., and Masdeu-Bultó, A.M. (2006) Eur. J. Inorg. Chem., 1067–1075. Meehan, N.J., Sandee, A.J., Reek, J.N.H., Kamer, P.C.J., van Leeuwen, P.W.N.M., and Poliakoff, M. (2000) Chem. Commun., 1497–1498. Licence, P., Kie, J., Sokolova, M., Ross, S.K., and Poliakoff, M. (2003) Green Chem., 5, 99–104. Chouchi, D., Gourgouillon, D., Courel, M., Vital, J., and Nunes da Ponte, M. (2001) Ind. Eng. Chem. Res., 40, 2551–2554. Burgener, M., Ferri, D., Grunwaldt, J.D., Mallat, T., and Baiker, A. (2005) J. Phys. Chem. B, 109, 16794– 16800. Burgener, M., Wirz, R., Mallat, T., and Baiker, A. (2004) J. Catal., 228, 152– 161. Hancu, D. and Beckman, E.J. (2001) Green Chem., 3, 80–86. Srinivas, P. and Mukhopadhyay, M. (1997) Ind. Eng. Chem. Res., 36, 2066–2074. Armbruster, U., Martin, A., Smejkal, Q., and Kosslik, H. (2004) Appl. Catal. A General, 265, 237–246. Olsen, M.H.N., Salomão, G.C., Drago, V., Fernandes, C., Horn, A., Jr., Cardozo Filho, L., and Antunes, O.A.C. (2005) J. Supercrit. Fluids, 34, 119–124. Jenzer, G., Mallat, T., Maciejewski, M., Eigenmann, F., and Baiker, A. (2001) Appl. Catal. A General, 208, 125–133. Nolen, S.A., Lu, J., Brown, J.S., Pollet, P., Eason, B.C., Grifﬁth, K.N., Glaeser, R., Bush, D., Lamb, D.R., Liotta, C.L., Eckert, C.A., Thiele, G.F., and Bartels, K.A. (2002) Ind. Eng. Chem. Res., 41, 316–323. Hancu, D., Green, J., and Beckman, E.J. (2002) Ind. Eng. Chem. Res., 41, 4466–4474.

58 Steele, A.M., Zhu, J., and Tsang, S.C. (2001) Catal. Lett., 73, 9–13. 59 Tsang, S.C., Zhu, J., Steele, A.M., and Meric, P. (2004) J. Catal., 226, 435–442. 60 Ciriminna, R., Campestrini, S., and Magliaro, M. (2004) Adv. Synth. Catal., 346, 231–236. 61 Maayan, G., Ganchegui, B., Leitner, W., and Neumann, R. (2006) Chem. Commun., 2230–2232. 62 Sato, T., Watanabe, A., Hiyoshi, N., Shirai, M., and Itoh, N. (2007) J. Supercrit. Fluids, 43, 295–302. 63 Kimmerle, B., Grunwaldt, J.D., and Baiker, A. (2007) Top. Catal., 44, 285–292. 64 Haider, P., Kimmerle, B., Krumeich, F., Kleist, W., Grunwaldt, J.D., and Baiker, A. (2008) Catal. Lett., 125, 169–176. 65 Jenzer, G., Sueur, D., Mallat, T., and Baiker, A. (2000) Chem. Commun., 2247–2248. 66 Jenzer, G., Mallat, T., and Baiker, A. (2001) Catal. Lett., 73, 5–8. 67 Jenzer, G., Schneider, M.S., Wandeler, R., Mallat, T., and Baiker, A. (2001) J. Catal., 199, 141–148. 68 Caravati, M., Meier, D.M., Grunwaldt, J.D., and Baiker, A. (2006) J. Catal., 240, 126–136. 69 González-Núñez, M.E., Mello, R., Olmos, A., Acerete, R., and Asensio, G. (2006) J. Org. Chem., 71, 1039–1042. 70 Burgener, M., Tyszewski, T., Ferri, D., Mallat, T., and Baiker, A. (2006) Appl. Catal. A General, 299, 66–72. 71 Hou, Z., Theyssen, N., and Leitner, W. (2007) Green Chem., 9, 127–132. 72 Hou, Z., Theyssen, N., Brinkmann, A., Klementiev, K.V., Grünert, W., Bühl, M., Schmidt, W., Spliethoff, B., Tesche, B., Weidenthaler, C., and Leitner, W. (2008) J. Catal., 258, 315–323. 73 Wang, X., Kawanami, H., Dapurkar, S.E., Venkataramanan, N.S., Chatterjee, M., Yokoyama, T., and Ikushima, Y. (2008) Appl. Catal. A General, 349, 86–90. 74 Klibanov, A.M. (1997) Trends Biotechnol., 15, 97–101. 75 Hobbs, H.R., Kirke, H.M., Poliakoff, M., and Thomas, N.R. (2007) Angew. Chem., Int. Ed., 46, 7860–7863.

29

30

2 Utilization of Dense Carbon Dioxide as an Inert Solvent for Chemical Syntheses 76 Isaacs, N.S. and Keating, N. (1992) J. Chem. Soc. Chem. Commun., 876– 877. 77 Ikushima, Y., Saito, N., and Arai, M. (1992) J. Phys. Chem., 96, 2293–2297. 78 Clifford, A.A., Pople, K., Gaskill, W.J., Bartle, K.D., and Rayner, C.M. (1997) Chem. Commun., 595–596. 79 Clifford, A.A., Pople, K., Gaskill, W.J., Bartle, K.D., and Rayner, C.M. (1998) J. Chem. Soc. Faraday Trans., 94, 1451–1456. 80 Reaves, J.T. and Roberts, C.B. (1999) Chem. Eng. Commun., 171, 117–134. 81 Lin, B. and Akgerman, A. (1999) Ind. Eng. Chem. Res., 38, 4525–4530. 82 Thompson, R.L., Gläser, R., Bush, D., Liotta, C.L., and Eckert, C.A. (1999) Ind. Eng. Chem. Res., 38, 4220–4225. 83 Qian, J., Timko, M.T., Allen, A.J., Russell, C.J., Winnik, B., Buckley, B., Steinfeld, J.I., and Tester, J.W. (2004) J. Am. Chem. Soc., 126, 5465–5474. 84 Subramaniam, B. and Clark, M.C. (1999) US Patent 5, 907,075. 85 Poliakoff, M., Swan, T.M., Tacke, T., Hitzler, M.G., Ross, S.K., Wieland, S., and Smail, F.R. (2001) US Patent 6, 303,840. 86 Shen, Z., McHugh, M.A., Xu, J., Belardi, J., Kilic, S., Mesiano, A., Bane, S., Karnikas, C., Beckman, E., and Enick, R. (2003) Polymer, 44, 1491– 1498. 87 Kilic, S., Michalik, S., Wang, Y., Johnson, J.K., Enick, R.M., and Beckman, E.J. (2007) Macromolecules, 40, 1332–1341. 88 Dolbier, W.R., Jr (1996) Chem. Rev., 96, 1557–1584. 89 Romack, T.J., Kipp, B.E., and DeSimone, J.M. (1995) Macromolecules, 28, 8432–8434. 90 Romack, T.J., Combes, J.R., and DeSimone, J.M. (1995) Macromolecules, 28, 1724–1726. 91 Giaconia, A., Scialdone, O., Apostolo, M., Filardo, G., and Galia, A. (2008) J. Polym. Sci. Part A Polym. Chem., 46, 257–266. 92 Charpentier, P.A., Kennedy, K.A., DeSimone, J.M., and Roberts, G.W. (1999) Macromolecules, 32, 5973–5975.

93 Charpentier, P.A., DeSimone, J.M., and Roberts, G.W. (2000) Ind. Eng. Chem. Res., 39, 4588–4596. 94 Saraf, M.K., Gerard, S., Wojcinski, L.M., II, Charpentier, P.A., DeSimone, J.M., and Roberts, G.W. (2002) Macromolecules, 35, 7976–7985. 95 Ahmed, T.S., DeSimone, J.M., and Roberts, G.W. (2004) Chem. Eng. Sci., 59, 5139–5144. 96 Galia, A., Caputo, G., Spadaro, G., and Filardo, G. (2002) Ind. Eng. Chem. Res., 41, 5934–5940. 97 Tai, H., Wang, W., Martin, R., Liu, J., Lester, E., Licence, P., Woods, H.M., and Howdle, S.M. (2005) Macromolecules, 38, 355–363. 98 Tai, H., Wang, W., and Howdle, S.M. (2005) Macromolecules, 38, 1542– 1545. 99 Galia, A., Giaconia, A., Scialdone, O., Apostolo, M., and Filardo, G. (2008) J. Polym. Sci. Part A Polym. Chem., 44, 2406–2418. 100 Mueller, P.A., Storti, G., Morbidelli, M., Costa, I., Galia, A., Scialdone, O., and Filardo, G. (2006) Macromolecules, 39, 6483–6488. 101 De Young, J.P., Romack, T.J., and DeSimone, J.M. (1997) Polym. Prepr., 38, 424–425. 102 Wood, C.D., Michel, U., Rollanda, J.P., and DeSimone, J.M. (2004) J. Fluorine Chem., 125, 1671–1676. 103 Michel, U., Resnick, P., Kipp, B., and DeSimone, J.M. (2003) Macromolecules, 36, 7107–7113. 104 Lousenberg, R.D. and Shoichet, M.S. (2000) Macromolecules, 33, 1682– 1685. 105 Baradie, B. and Shoichet, M.S. (2002) Macromolecules, 35, 3569–3575. 106 Ahmed, T.S., DeSimone, J.M., and Roberts, G.W. (2006) Macromolecules, 39, 15–18. 107 Tai, H., Wang, W., and Howdle, S.M. (2005) Macromolecules, 38, 9135–9142. 108 Beginn, U., Najjar, R., Ellmann, J., Vinokur, R., Martin, R., and Moller, M. (2006) J. Polym. Sci. Part A Polym. Chem., 44, 1299–1316. 109 Hile, D.D. and Pishko, M.V. (1999) Macromol. Rapid Commun., 20, 511–514.

References 110 Mazarro, R., de Lucas, A., Gracia, I., and Rodríguez, J.F. (2008) J. Biomed. Mater. Res. B Appl. Biomater., 85B, 196–203. 111 Hile, D.D. and Pishko, M.V. (2001) J. Polym. Sci. Part A Polym. Chem., 39, 562–570. 112 Bratton, D., Brown, M., and Howdle, S.M. (2003) Macromolecules, 36, 5908–5911.

113 Bratton, D., Brown, M., and Howdle, S.M. (2005) J. Polym. Sci. Part A Polym. Chem., 43, 6573–6585. 114 Ganapathy, H.S., Hwang, H.S., Jeong, Y.T., Lee, W.K., and Lim, K.T. (2007) Eur. Polym. J., 43, 119–126. 115 Loeker, F.C., Duxbury, C.J., Kumar, R., Gao, W., Gross, R.A., and Howdle, S.M. (2004) Macromolecules, 37, 2450–2453.

31

33

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations Ivan A. Berg, Daniel Kockelkorn, W. Hugo Ramos-Vera, Rafael Say, Jan Zarzycki, and Georg Fuchs

3.1 Introduction

Animals, fungi, and many microbes, that require organic compounds as a carbon source are referred to as heterotrophs. They oxidize such compounds to carbon dioxide. Autotrophs (= self-nourishing organisms), in contrast, are able to assimilate inorganic carbon to build all cell constituents, and are responsible for the primary production on Earth. Autotrophic carbon ﬁxation, therefore, is essential for life, representing the most important synthetic process in biology (2 × 1011 tons of CO2 are ﬁxed annually). This capability is a characteristic feature of prokaryotes, and only by endosymbiosis was it conferred on eukaryotic cells [1]. Chloroplasts, as well as many autotrophic bacteria, use the Calvin–Benson– Bassham (CBB) cycle for CO2 ﬁxation [2]. The later discovery of a second autotrophic pathway, the reductive citric acid cycle [3], was unexpected and for a long time this was not accepted by the scientiﬁc community as it seemingly contradicted a biological dogma of that time, namely the biochemical unity of life. Yet, today, four other fundamentally different autotrophic processes have been identiﬁed in prokaryotes that allow cellular building blocks to be synthesized from CO2. In general terms, the endergonic assimilation of CO2 into cellular building blocks requires reducing equivalents and an input of energy. It can be formulated as: CO2 + 4 [H] + nATP (CH2O) + H2O + nADP + nPi

(3.1)

where (CH2O) represents cell carbon at an average oxidation state zero, as in carbohydrates. Reducing power is provided by inorganic compounds such as water, hydrogen gas, reduced sulfur compounds, or ammonia. Likewise, energy is provided by photosynthesis or by respiration – that is, by the exergonic reduction of oxidized inorganic compounds such as oxygen, nitrate, or sulfate that function as “terminal electron acceptors” in respiration. Therefore, autotrophic CO2 ﬁxation occurs both in the light and in the dark, and in both aerobic and anaerobic environments. Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

34

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

The usage of energy produced through the oxidation of inorganic compounds is termed chemosynthesis; this is in contrast to photosynthesis, where light energy is used. Although, nowadays, phototrophic green plants are responsible for the main primary production, they were preceded by chemotrophs in evolution, and the niches where chemoautotrophs are responsible for primary CO2 production are relics today. An example of such an ecosystem is the deep-sea vents on the dark seaﬂoor, where the energy and carbon sources, mainly H2, H2S, CO and CO2, are supplied by magma degassing and from high-temperature reactions between seawater and rocks [4]. Such ecosystems are ultimately driven by the energy of sunlight, because the energy-producing respiration processes of the organisms depend on oxidized inorganic compounds (O2 or sulfate). These electron acceptors of respiration are the products of oxygenic photosynthesis (O2), or of the oxidation of reduced sulfur compounds with photosynthetically produced oxygen. The question is therefore, what are the principal requirements of an autotrophic carbon-ﬁxation mechanism? An organic molecule serves as a CO2 acceptor molecule, which becomes carboxylated by a carboxylase enzyme. This CO2 acceptor molecule needs to be regenerated in a reductive autocatalytic cycle. The product that can be drained off from such a metabolic cycle should be a central cellular metabolite, from which all cellular building blocks for polymers can be derived; examples of such central metabolites are acetyl-CoA, pyruvate, oxaloacetate, 2-oxoglutarate, phosphoenolpyruvate, and 3-phosphoglycerate. Importantly, the intermediates should not be toxic to the cell. The irreversible steps of the pathway are driven by ATP hydrolysis, while the reduction steps are driven by low-potential reduced coenzymes. This chapter summarizes the current knowledge of CO2 ﬁxation pathways, and emphasizes the characteristic features that determine their distribution in the modern biosphere.

3.2 The Mechanisms of CO2 Fixation

Presently, six CO2 ﬁxation pathways are known; these are listed and described in the order of their discovery. 3.2.1 The Calvin–Benson–Bassham (CBB) Cycle

This cycle represents the quantitatively most important CO2 ﬁxation pathway in Nature. It is found in most aerobic autotrophic organisms, ranging from diverse photosynthetic and chemolithoautotrophic bacteria to chloroplasts of eukaryotic algae and higher plants [5]. It is centered around carbohydrates, with ribulose 1,5-bisphosphate being the CO2 acceptor (Figure 3.1). CO2 (and oxygen; see below) adds to C2 of this molecule in its enediol form. Basically, only two key enzymes are required in addition to the enzymatic outﬁt of most cells, namely the carboxylase ribulose 1,5-bisphosphate carboxylase/oxy-

3.2 The Mechanisms of CO2 Fixation

CH2OH C O 5 x Glyceraldehyde 3-phosphate

CHOH

Glyceraldehyde 3-phosphate

CHOH CH2O P 3 x Ribulose-5-phosphate 3 ATP

6 ADP + 6 Pi

1 6 NADPH

3 ADP

COOH

CH2O P

CHOH

C O

CH2O P

CHOH

6 x 3-Phosphoglycerate

CHOH CH2O P

6 ATP

3 CO2

2

3 x Ribulose 1,5-bisphosphate 3 H2O

Figure 3.1 The Calvin–Benson–Bassham (CBB) cycle. 1,5-bisphosphate carboxylase/oxygenase (RubisCO).

, phosphoribulokinase;

, ribulose

genase (RubisCO), and phosphoribulokinase that is required for regeneration of the CO2 acceptor. Ribulose 1,5-bisphosphate carboxylation and hydrolysis of the resulting C6 intermediate leads to two molecules of 3-phosphoglycerate, which are further reduced with NAD(P)H to glyceraldehyde 3-phosphate, driven by ATP hydrolysis. The regeneration part of the cycle consists of the interconversion of triosephosphates via various sugar phosphates to ribulose 5-phosphate, which is phosphorylated by phosphoribulokinase to ribulose 1,5-bisphosphate. Three rounds of the cycle generate one molecule of triosephosphate. The energy demand of the CBB cycle is high, since nine ATP are required to form one triosephosphate molecule from three molecules of CO2 (see Table 3.1 for a comparison with other cycles). The key CBB cycle enzyme, RubisCO, is the most abundant protein in the world [8], as it can comprise up to 50% of the total soluble protein in the chloroplasts or in bacteria using this cycle. This fact is a consequence of the notorious catalytic inefﬁciency of RubisCO, that is, a low afﬁnity for CO2, a slow catalytic turnover rate, and a wasteful oxygenase side reaction responsible for photorespiration, resulting in a futile cleavage of the substrate to form phosphoglycolate as a side product. However, the CBB cycle enzymes are oxygen-insensitive and can easily be controlled, because the whole pathway is separated from

35

Comparison of autotrophic CO2-ﬁxation pathways.

Pathway

Amount of ATP for synthesis of one triose phosphate

CO2-ﬁxing enzymes

Active “CO2” species

CO2 ﬁxation products which may be used for biosynthesis

Key enzymes 36

Reductants for synthesis of one triose phosphate

Calvin–Benson–Bassham cycle (reductive pentose phosphate cycle)

9

6 NAD(P)H

RubisCO

CO2

3-Phosphoglycerate

RubisCO; Phosphoribulokinase

Reductive citric acid cycle

5

3 NAD(P)H, 1 unknown donora, 2 ferredoxin

2-Oxoglutarate synthase; Isocitrate dehydrogenaseb; Pyruvate synthase; PEP carboxylase

CO2; CO2; CO2; HCO3−

Acetyl-CoA, pyruvate, PEP, oxaloacetate, succinyl-CoA, 2-oxoglutarate

2-Oxoglutarate synthase, ATP-citrate lyase

Reductive acetyl-CoA pathway

4–5

3 NAD(P)H, 2–3 ferredoxin, 1 H2 (in methanogens)

Acetyl-CoA synthase/ CO dehydrogenase, formate dehydrogenase, pyruvate synthase

CO2

Acetyl-CoA, pyruvate

Acetyl-CoA synthase/CO dehydrogenase, enzymes reducing CO2 to methyltetrahydropterin

7 NAD(P)H, but 1 FAD is reduced in the cycle

Acetyl-CoA/propionylCoA carboxylase

HCO3−

Acetyl-CoA, pyruvate, succinyl-CoA

Malonyl-CoA reductase, propionyl-CoA synthase, malyl-CoA lyase

3-Hydroxypropionate/ malyl-CoA cycle

10

3-Hydroxypropionate/ 4-hydroxybutyrate cycle

9

6 NAD(P)H

Acetyl-CoA/propionylCoA carboxylase

HCO3−

Acetyl-CoA, succinyl-CoA

Acetyl-CoA/propionyl-CoA carboxylasec, enzymes reducing malonyl-CoA to propionyl-CoA, methylmalonyl-CoA mutasec, 4-hydroxybutyryl-CoA dehydratase

Dicarboxylate/ 4-hydroxybutyrate cycled

8

2 NAD(P)H, 3 ferredoxin, 1 unknown donor

Pyruvate synthase, PEP carboxylase

CO2; HCO3−

Acetyl-CoA, pyruvate, PEP, oxaloacetate, succinyl-CoA

4-Hydroxybutyryl-CoA dehydratase

a) b) c) d)

NADH in Hydrogenobacter thermophilus [6]. Biotin-dependent 2-oxoglutarate carboxylase in Hydrogenobacter thermophilus [7]. The presence of these enzymes is usual for Bacteria, but not Archaea, where this pathway was discovered. As studied in Ignicoccus hospitalis.

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

Table 3.1

3.2 The Mechanisms of CO2 Fixation

most of the central metabolism of the cell, and sugar phosphates are common metabolites. Up to now, four forms of RubisCO have been recognized, with forms I–III being bona ﬁde RubisCO, and form IV referred to as RubisCO-like protein (RLP). First identiﬁed in the phototrophic green sulfur bacterium Chlorobium tepidum [9], RLPs are found in a wide variety of prokaryotes. Although RLPs are structurally related to the bona ﬁde RubisCO, they do not function as RubisCO enzymes but catalyze a reaction in sulfur metabolism [5]. Form III RubisCO is conﬁned to Archaea, and its role in their metabolism is at issue [5, 10–12]. Interestingly, RubisCO was found in many Archaea that were not reported to be able to grow autotrophically. On the contrary, RubisCO has also been found in Archaea that grow autotrophically but are considered to use the reductive acetyl-CoA pathway for CO2 ﬁxation. Hence, they contain at the same time RubisCO activity and the key enzyme activities of the reductive acetyl-CoA pathway. Notably, phosphoribulokinase, the second key enzyme of the CBB cycle, was not detected in Archaea, and the cycle cannot be closed yet. Although, the possibility to synthesize the substrate of RubisCO, ribulose 1,5-bisphosphate, was shown in some Archaea containing RubisCO form III [10, 11], such a pathway cannot be closed in a cycle, and the understanding of the function of this enzyme is still elusive; however, its participation in AMP metabolism has been proposed recently [11]. The ﬁnding that genomes of three methanogens contain a gene for phosphoribulokinase in addition to a form III RubisCO gene [12] reinforces the possibility that the CBB cycle may function in some Archaea in addition to the reductive acetyl-CoA pathway. A recent phylogenetic analysis showed a possible archaeal origin of both RubisCO and RLP, with form III proteins from Methanomicrobia being the likely precursors of all modern RubisCO and RLP lineages [5]. This is in agreement with the proposed late appearance of the CBB cycle in evolution [13–15]. Its immediate success is not obvious, when considering the adverse characteristics of its key enzyme, RubisCO. However, the oxygen-insensitivity of this carboxylase may have been decisive, despite its oxygenase side activity. The CBB cycle likely evolved before the oxygen concentration in the atmosphere increased to modern high values. 3.2.2 The Reductive Citric Acid Cycle (Arnon–Buchanan Cycle)

This cyclic pathway was ﬁrst proposed for Chlorobium sp. that use an anoxygenic photosynthesis for energy supply [3]. It reverses the reactions of the oxidative citric acid cycle (Krebs cycle) and forms acetyl-CoA from two CO2 (Figure 3.2). Three modiﬁcations of the conventional oxidative citric acid cycle are needed, which substitute irreversible enzyme steps. Succinate dehydrogenase is replaced by fumarate reductase, 2-oxoglutarate dehydrogenase by ferredoxin-dependent 2-oxoglutarate oxidoreductase (2-oxoglutarate synthase), and citrate synthase by ATP-citrate lyase [3, 16]; it should be noted that the carboxylases of the cycle catalyze the reductive carboxylation reactions. There are variants of the ATP-driven cleavage of citrate as well as of isocitrate formation [7]. The reductive citric acid

37

38

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

CO2

COOH C

O 2 Fdred

CH3

4 COSCoA CH3 Acetyl-CoA CoA

Pyruvate

COOH COOH

ATP

ADP+Pi

C O

1

CH2 HO

C COOH

CH2

CH2

COOH

COOH

Oxaloacetate

Citrate

COOH CHOH

COSCoA

HC COOH

CH2 CH2 COOH Succinyl-CoA

COOH

CO2

2

C O

CO2

CH2 2 Fdred CoA

3

CH2 COOH Isocitrate

NAD(P)H

CH2 COOH 2-Oxoglutarate

Figure 3.2 Reductive citric acid cycle. , ATP-citrate lyase; , 2-oxoglutarate : ferredoxin oxidoreductase (2-oxoglutarate synthase); , isocitrate dehydrogenase; , pyruvate : ferredoxin oxidoreductase (pyruvate synthase). Fdred = reduced ferredoxin.

cycle (in Chlorobium) requires ﬁve ATP to form triosephosphates from three CO2, and depends on oxygen-sensitive carboxylases that use reduced ferredoxin as the electron donor for the reductive carboxylation of the CoA thioesters of acetate and succinate (see Table 3.1). This small iron–sulfur protein is also oxygen-sensitive. The pathway may be used even in the reversed oxidative direction for the oxidation of acetyl-CoA in one and the same organisms (Desulfobacter sp.) [17]. The use of ATP citrate lyase reaction (Equation 3.2), instead of the irreversible citrate synthase reaction, allows the generation of a second molecule of ATP in the oxidative cycle, in addition to ATP synthesis via succinate thiokinase. It is not known, which mechanism determines the direction of the cycle in Desulfobacter – that is, whether a compound is preferentially oxidized or assimilated. Acetyl-CoA + Oxaloacetate + ADP + Pi → Citrate + CoA + ATP

(3.2)

This pathway involves oxygen-sensitive enzymes and electron carriers, and therefore is found in anaerobic or microaerobic bacteria, such as anaerobic sulfate reducers

3.2 The Mechanisms of CO2 Fixation

(e.g., Desulfobacter sp., Deltaproteobacteria), microaerobic sulfur oxidizers (Epsilonproteobacteria), anaerobic phototrophic green sulfur bacteria (Chlorobium), and hydrogen-oxidizing microaerobic Aquiﬁcales (Aquifex, Hydrogenobacter) [3, 7, 18, 19]. The functioning of this cycle in anaerobic Crenarchaeota (e.g., Thermoproteus sp.) [20] is now questioned, as all enzymes of the dicarboxylate/4-hydroxybutyrate cycle have been found there (see below). The low ATP requirement, as well as the moderate oxygen sensitivity of some of its enzymes, makes the reductive citric acid cycle a suitable pathway mainly for anaerobes, but also for some microaerobes. 3.2.3 The Reductive Acetyl-CoA Pathway (Wood–Ljungdahl Pathway)

This is a noncyclic pathway that also results in the ﬁxation of two molecules of CO2 to form acetyl-CoA. It was elucidated by Wood, Ljungdahl, Thauer and others as a pathway which is used by acetogenic bacteria to synthesize acetate from CO2 in their energy metabolism [21]. The acetyl-CoA pathway resembles the Monsanto process of acetate synthesis from CO and methanol, with one molecule of CO2 being reduced to the level of methyltetrahydropterin, while another CO2 molecule is reduced to the level of carbon monoxide in the reaction catalyzed by the nickeldependent carbon monoxide dehydrogenase (Figure 3.3). This enzyme also acts as acetyl-CoA synthase, by accepting the methyl group from a methylated corrinoid protein (the methyl group being derived from methyltetrahydropterin), combining it with the carbon monoxide group to form an enzyme-bound acetyl group, and then releasing this group thiolytically with coenzyme A to form acetyl-CoA. This key enzyme CO dehydrogenase/acetyl-CoA synthase has probably common roots in all prokaryotes, in contrast to those enzymes involved in the formation of methyltetrahydropterin from CO2. There are many variants of the pathway which differ in the use of coenzymes and electron carriers; the pathway can also be reversed and used for the oxidation of acetyl-CoA, instead of the citric acid cycle [22]. The reductive acetyl-CoA pathway is unique in several aspects. For example, the pathway makes extensive use of coenzymes (tetrahydropterin, cobalamin) and of

6 [H]

CO2

[CH3 ] 2 H2O 2 [H]

CO2

1

1

CH3 COSCoA

CoASH

[CO ] H2O

Acetyl-CoA

CO2

2 Fdred

2

CH3 CO COOH Pyruvate

Figure 3.3 Reductive acetyl-CoA pathway. , pyruvate : ferredoxin oxidoreductase.

, CO dehydrogenase/acetyl-CoA-synthase;

39

40

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

metals (Mo or W, Co, Ni, Fe–S centers). It also enables CO, formaldehyde, methanol, methylamine, or methylmercaptane to be assimilated. Many of the onecarbon units react spontaneously with the cofactors and prosthetic groups of this pathway. This pathway depends on strict anoxic conditions, as some of its enzymes – notably CO dehydrogenase/acetyl-CoA synthase – are highly oxygensensitive. It also requires the above-mentioned metals, which are water-soluble preferentially in the reduced oxidation state – that is, under anoxic conditions. The process can be simulated in the laboratory to produce not only acetylthioesters but also derived products by simply incubating CO, H2 and H2S or methylmercaptane with Ni and Fe salts; these inorganic metals form mixed Ni–Fesulﬁdes that act as catalysts [23, 24]. Among the CO2 ﬁxation pathways, this pathway has the lowest energetic costs, requiring only four ATPs per triose phosphate (see Table 3.1). However, its requirements for metals, cofactors, anaerobiosis are among the highest, which causes it to be restricted to a limited number of ecological niches. The reductive acetyl-CoA pathway is the favored autotrophic pathway in strict anaerobes. Notably, it is found in those bacteria that encounter traces of the abovementioned gases and one-carbon compounds in their environment. Acetyl-CoA is assimilated into cell material as described for the reductive citric acid cycle, and can simultaneously serve as an energy source for ATP synthesis during its conversion to acetate. This pathway is found in acetogenic bacteria and methanogenic Archaea, in sulfate-reducing bacteria (Desulfobacterium sp. (Deltaproteobacteria), Desulfotomaculum sp. (relatives of Clostridia), Archaeoglobus (Euryarchaeota)), as well as in anaerobic ammonia-oxidizing bacteria [13, 25, 26]. 3.2.4 The 3-Hydroxypropionate/Malyl-CoA Cycle

This pathway results in the ﬁxation of three molecules of bicarbonate, and forms pyruvate as the central carbon precursor molecule. The main CO2-ﬁxing enzyme is acetyl-CoA/propionyl-CoA carboxylase. The pathway can be divided into two metabolic cycles (Figure 3.4). In the ﬁrst cycle, acetyl-CoA is carboxylated to malonyl-CoA, which is subsequently reduced and converted into propionyl-CoA via 3-hydroxypropionate as a free intermediate. Propionyl-CoA is carboxylated to methylmalonyl-CoA, which is subsequently converted to succinyl-CoA; the latter is then used to activate L-malate by succinylCoA:L-malate coenzyme A transferase, which forms L-malyl-CoA and succinate. Succinate is oxidized to L-malate via conventional steps. L-Malyl-CoA, the second characteristic intermediate of this cycle, is cleaved by L-malyl-CoA/β-methylmalylCoA lyase, thus regenerating the starting molecule acetyl-CoA and releasing glyoxylate as a ﬁrst carbon-ﬁxation product [27]. Glyoxylate is an unconventional cell carbon precursor that requires special enzymes to be used in its biosynthesis; a second cycle serves as the glyoxylate assimilation pathway [28]. Glyoxylate is combined with propionyl-CoA to β-

3.2 The Mechanisms of CO2 Fixation Glyoxylate

41

O

H

(S)-Malyl-CoA CoA-S

Pyruvate

CoA-S

COOH O

COOH

H3C

COOH O CH3

O

2 Acetyl-CoA O

OH

2 HCO3¯ CoA-S

OH

(S)-Malate H2O

CoA-S

Succinyl-CoA 2 Propionyl-CoA (S)-Methylmalonyl-CoA

CH3

1st Cycle

HCO3¯

S-CoA O

H2O CH3

HOOC

S-CoA

ADP + Pi ATP

COOH

CH3

Mesaconyl-C1-CoA

S-CoA OH

O

O

CH3

Mesaconyl-C4-CoA

2 CoA 2 NADPH + 2 H+ + 2 ATP 2 NADP+ + 2 AMP + 2 PPi

S-CoA

HOOC

O

HOOC

2 3-Hydroxypropionate

CH3

H2O

COOH

HO COOH

OH

CoA-S

CoA-S

4 NADPH + 4 H+ 4 NADP+ 2 CoA

2 [H] O

O

COOH

2 Malonyl-CoAO

CoA

OH

2 ATP 2 ADP + 2 Pi

COOH

HOOC

(S)-Citramalyl-CoA H3C

O

β-Methylmalyl-CoA

H HOOC O

Glyoxylate

2nd Cycle

Figure 3.4 3-Hydroxypropionate/malyl-CoA cycle, as studied and proposed in Chloroﬂexus aurantiacus.

methylmalyl-CoA, which in turn is converted via mesaconyl-CoA to citramalylCoA. The latter is cleaved into pyruvate and acetyl-CoA. Acetyl-CoA conversion to propionyl-CoA occurs as described above, closing the second, glyoxylate assimilation cycle. The distribution of this pathway is quite limited; to date, it has only been observed in the phototrophic bacterium Chloroﬂexus aurantiacus, a member of the green nonsulfur bacteria [27]. Although the genomes of closely related bacteria, Chloroﬂexus aggregans and Roseiﬂexus spp., contain the genes for this pathway [29], they appear to be unable to grow autotrophically. Interestingly, another green nonsulfur bacterium, Oscillochloris sp., uses the CBB cycle for autotrophy [30]. It seems reasonable to suggest that the 3-hydroxypropionate/malyl-CoA cycle evolved in the Chloroﬂexaceae, and has close connections with heterotrophic metabolism. The cycle allows mixotrophic growth by assimilating fermentation products such as acetate, propionate, or succinate, which makes it suitable for bacteria living in habitats where such fermentation products may occur. Indeed, the preferred growth mode of Chloroﬂexus sp. is photoheterotrophy.

42

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

An interesting feature of this cycle is the presence of a number of bi-/multifunctional enzymes: bifuctional malonyl-CoA reductase, which catalyzes two consecutive reductase steps leading to 3-hydroxypropionate; trifunctional propionyl-CoA synthase, a fusion protein converting 3-hydroxypropionate to propionyl-CoA in three steps; and trifunctional malyl-CoA/β-methylmalyl-CoA/citramalyl-CoA lyase, which not only cleaves L-malyl-CoA into acetyl-CoA and glyoxylate (last reaction of the ﬁrst cycle, Figure 3.4), but also synthesizes β-methylmalyl-CoA from glyoxylate and propionyl-CoA and splits L-citramalyl-CoA into pyruvate and acetylCoA, thus opening and closing the glyoxylate assimilation cycle [31–33]. The energy costs of the 3-hydroxypropionate/malyl-CoA cycle are high, with ten ATP required per triose phosphate (see Table 3.1). However, bicarbonate rather than CO2 is the actual inorganic carbon species used by acetyl-CoA/propionyl-CoA carboxylase (this is discussed in Chapter 4). Moreover, as this enzyme is virtually irreversible and has a high afﬁnity for bicarbonate, this cycle is expensive although kinetically effective. 3.2.5 The 3-Hydroxypropionate/4-Hydroxybutyrate Cycle

In this cycle, one molecule of acetyl-CoA is formed from two molecules of bicarbonate (Figure 3.5). The key carboxylating enzyme is the bifunctional biotindependent acetyl-CoA/propionyl-CoA carboxylase. In Bacteria and Eukarya, acetyl-CoA carboxylase catalyzes the ﬁrst step of fatty acid biosynthesis. However, Archaea do not contain fatty acids in their lipids, and acetyl-CoA carboxylase cannot serve as the key enzyme of fatty acid synthesis; rather, it is responsible for autotrophy. The product of acetyl-CoA carboxylase reaction, malonyl-CoA, is reduced via malonate semialdehyde to 3-hydroxypropionate, which is further reductively converted to propionyl-CoA. Propionyl-CoA is carboxylated to (S)-methylmalonyl-CoA by the same carboxylase. (S)-Methylmalonyl-CoA is isomerized to (R)-methylmalonyl-CoA, followed by carbon rearrangement to succinyl-CoA by coenzyme B12dependent methylmalonyl-CoA mutase. Succinyl-CoA is further reduced to succinate semialdehyde and then to 4-hydroxybutyrate. The latter compound is converted into two acetyl-CoA molecules via 4-hydroxybutyryl-CoA dehydratase, a key enzyme of the pathway. 4-Hydroxybutyryl-CoA dehydratase is a [4Fe-4S] cluster and FAD-containing enzyme that catalyzes the elimination of water from 4-hydroxybutyryl-CoA by a ketyl radical mechanism to yield crotonyl-CoA [34]. Conversion of the latter into two molecules of acetyl-CoA proceeds via normal βoxidation steps. Hence, the 3-hydroxypropionate/4-hydroxybutyrate cycle (as illustrated in Figure 3.5) can be divided into two parts. In the ﬁrst part, acetyl-CoA and two bicarbonate molecules are transformed to succinyl-CoA, while in the second part succinyl-CoA is converted to two acetyl-CoA molecules. The 3-hydroxypropionate/4-hydroxybutyrate cycle functions in autotrophic Sulfolobales (Crenarchaeota) [35–37]. These are extreme thermoacidophiles from volcanic areas which grow best at a pH of about 2 and temperatures of 60 to 90 °C.

3.2 The Mechanisms of CO2 Fixation

43

CoAS

acetyl-CoA

O

HCO3

ADP+Pi

CoAS acetoacetyl-CoA

CoAS NADH+H NAD

acetyl-CoA CoASH

+

+

O

malonyl-CoA

O ATP

CoAS

O

O O

CoAS

NADPH+H+

OH

NADP++CoASH

malonate semialdehyde

O

O

OH

(S)-3-hydroxybutyryl-CoA

NADPH+H+

O

H2O

OH NADP

CoAS

3-hydroxypropionate

+

O

O

crotonyl-CoA

OH OH ATP+CoASH

H2 O

OH

AMP+PPi

CoAS O

3-hydroxypropionyl-CoA

4-hydroxybutyryl-CoA

O

OH

AMP+PPi ATP+CoASH

acryloyl-CoA

HO 4-hydroxybutyrate NADP

SCoA

+

NADPH+H+

H 2O

O

O

NADPH+H+

O HO O

NADP+

propionyl-CoA

succinic semialdehyde

NADP

O

ATP +

NADPH+H+

O HO

SCoA O succinyl-CoA

O

SCoA

OH O

HCO3

-

ADP+Pi

SCoA

(S)-methylmalonyl-CoA

Figure 3.5

SCoA

OH

3-Hydroxypropionate/4-hydroxybutyrate cycle, as studied in Metallosphaera sedula.

Most Sulfolobales can grow chemoautotrophically by oxidizing sulfur, pyrite, or hydrogen under microaerobic conditions. This cycle was also proposed for the mesophilic ammonia-oxidizing symbiotic Cenarchaeum symbiosum, a member of the “marine group”-1 Crenarchaeota which are abundant in the sea [37]. The enzymes of the 3-hydroxypropionate/4-hydroxybutyrate cycle are oxygen-tolerant. Although 4-hydroxybutyryl-CoA dehydratase is inactivated by oxygen, it may be sufﬁciently stable at low oxygen tensions to maintain activity, especially in the protected environment of the cell. Therefore, the 3-hydroxypropionate/4-hydroxybutyrate cycle ﬁts well to the lifestyle of (micro)aerobic Crenarchaeota. The active “CO2” species in the 3-hydroxypropionate/4-hydroxybutyrate cycle is bicarbonate (see Table 3.1). The use of bicarbonate as a substrate may be advantageous for organisms using this cycle in comparison with, for example, the CBB

44

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

cycle. Metallosphaera sedula acetyl-CoA/propionyl-CoA carboxylase has a turnover number at 65 °C of 28 s−1 [38]. The Km value of RubisCO for dissolved CO2 (average 0.05 mM, but it may be as high as 0.34 mM in some marine cyanobacteria [39]) is lower than or even comparable to the apparent Km value of archaeal acetyl-CoA/ propionyl-CoA carboxylase for bicarbonate (0.3 mM) [38]. However, it should be noted that the CO2 concentration in water equilibrated with air (1 atm pressure, 20 °C, 370 ppm CO2) is 0.012 mM at pH 7.4 (assumed intracellular pH), and that of bicarbonate is 0.26 mM. The concentration of bicarbonate will be even higher at pH 7.8–8.2 of sea water (pKapp. CO2 HCO3− = 6.3, 20 °C). Thus, the acetyl-CoA carboxylase afﬁnity for the active “CO2” species appears not to be inferior to that of RubisCO. In addition, in volcanic gases and in animal hosts, the CO2 concentration is much higher than in air. It should also be noted that carboxyphosphate, which is used in these biotin-dependent carboxylase reactions, is an attractive model for carbon ﬁxation during chemoevolution. Although the 3-hydroxypropionate part of this cycle resembles the ﬁrst part of the 3-hydroxypropionate/malyl-CoA cycle functioning in Chloroﬂexus, the enzymes used to synthesize propionyl-CoA from malonyl-CoA are not homologous, although the intermediates are the same [31, 32, 37]. Therefore, this pathway appears to have evolved independently in Sulfolobales and in the Chloroﬂexaceae, representing an interesting example of convergent evolution. The energy requirement of the 3-hydroxypropionate/4-hydroxybutyrate cycle is nine ATPs for one triosephosphate synthesized (see Table 3.1). 3.2.6 The Dicarboxylate/4-Hydroxybutyrate Cycle

This cycle resembles the 3-hydroxypropionate/4-hydroxybutyrate cycle, but with pyruvate : ferredoxin oxidoreductase (pyruvate synthase) and phosphoenolpyruvate (PEP) carboxylase as the carboxylating enzymes (Figure 3.6). The dicarboxylate/4-hydroxybutyrate cycle starts from acetyl-CoA, which is reductively carboxylated to pyruvate. Pyruvate is converted to PEP and then carboxylated to oxaloacetate. The latter is reduced to succinyl-CoA by the reactions of an incomplete reductive citric acid cycle. Succinyl-CoA is reduced to 4-hydroxybutyrate, the subsequent conversion of which into two acetyl-CoA molecules proceeds in the same way as in the 3-hydroxypropionate/4-hydroxybutyrate cycle. The cycle can be divided into part 1 transforming acetyl-CoA, one CO2 and one bicarbonate to succinyl-CoA via pyruvate, PEP, and oxaloacetate, and part 2 converting succinyl-CoA via 4-hydroxybutyrate into two molecules of acetyl-CoA. This cycle was shown to function in Ignicoccus hospitalis, an anaerobic autotrophic hyperthermophilic Archaeum (Desulfurococcales) [40]. Moreover, this pathway functions in Thermoproteus neutrophilus (Thermoproteales), where the reductive citric acid cycle was earlier assumed to operate, but was later disproved (W.H. Ramos-Vera et al., unpublished results). The active “CO2” species in the dicarboxylate/4-hydroxybutyrate cycle are CO2 as cosubstrate for pyruvate synthase, and bicarbonate as cosubstrate for PEP

3.2 The Mechanisms of CO2 Fixation

45

CoAS

acetyl-CoA

O

CO2 CoAS

2 Fdox+CoASH

O

acetyl-CoA

CoAS

2 Fdred

CoASH

NADH+H+

O NAD

HO

O

O

O

pyruvate

ATP+H2O

acetoacetyl-CoA

+

Pi+AMP

HO

CoAS O

O

O

OH

(S)-3-hydroxybutyryl-CoA

P

PEP

-

HCO3 Pi

O

H2 O

O

HO

CoAS crotonyl-CoA

OH

O

O

oxaloacetate NADH+H+

H2 O NAD

OH

OH O

CoAS O

HO

4-hydroxybutyryl-CoA

H2 O

O

OH HO

HO

fumarate

4-hydroxybutyrate

O

O NAD(P)+

O

O

NAD(P)H+H+

HO

HO O

succinic semialdehyde

HO

2 MVox

OH

O succinate SCoA

O succinyl-CoA

OH

2 MVred

O 2 MVox+CoASH 2 MVred

Figure 3.6

OH

(S)-malate

O

AMP+PPi ATP+CoASH

+

ATP+CoASH ADP+Pi

Dicarboxylate/4-hydroxybutyrate cycle, as studied in Ignicoccus hospitalis.

carboxylase (see Table 3.1). However, as the organisms bearing this cycle live in volcanic areas with a high ambient CO2 partial pressure, the afﬁnity of the carboxylases for CO2 or bicarbonate may be less critical than in the case of the CBB cycle. Triose phosphate formation in this cycle requires eight ATP equivalents (see Table 3.1). A comparison to the crenarchaeal 3-hydroxypropionate/ 4-hydroxybutyrate cycle reveals that the dicarboxylate/4-hydroxybutyrate cycle preferentially uses reduced ferredoxin instead of NAD(P)H as electron donor in the reduction steps. The oxygen sensitivity of some of its enzymes and electron carriers restricts this cycle to anaerobic organisms. Indeed, Thermoproteus and Ignicoccus species grow as strict anaerobes by reducing elemental sulfur with hydrogen gas.

46

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

3.3 Rules to Explain the Diversity

The six autotrophic pathways differ in various aspects such as: the use of CO2 or bicarbonate; the afﬁnity of the key carboxylase(s) for CO2 or bicarbonate; the redox potential of the reduction steps; the coenzymes involved and the natural inorganic substrates used for the reduction of the electron carriers; the ATP requirement; the requirements for metals (Fe, Co, Ni, Mo); the oxygen sensitivity of the enzymes; the interconnection of the carbon metabolism and the energy metabolism; the regulation mechanisms; the capability of concomitant assimilation of common low-molecular-mass compounds, such as fermentation products; the possibility to use the reverse reactions of the pathways for the complete oxidation of organic compounds; the evolutionary origin of the machinery; and the distance to the presumed chemoevolutionary scenario. These criteria determine the distribution of the pathways in autotrophic organisms in different habitats, in addition to their phylogeny. The pathways also differ in 12C/13C isotope discrimination, a fact that has long been neglected and which has an impact on the interpretation of carbon isotope fractionation data and geological records. The distribution of the CO2 ﬁxation pathways in the different lineages of Prokaryotes depends on the individual characteristic features of the pathways and enzymes, which determine the adaptation of their bearer to its natural environment. For instance, highly oxygensensitive enzymes are useless for aerobic organisms, while pathways with high energy demands are not appropriate for strict anaerobes that live at the expense of an energy metabolism that is close to the thermodynamic equilibrium. Furthermore, organisms belonging to different phylogenetic groups have inherited a complex enzyme outﬁt that cannot be changed suddenly at will. Different ecological strategies may also be applied in the different niches and by the different prokaryotic groups [41]. For instance, some autotrophic microorganisms settle in environments in which the fermentation products of other microorganisms frequently occur and can be used as an additional carbon source. The efﬁciency of different CO2 ﬁxation pathways is normally judged by comparing their energy demands. However, such a simple comparison is not always justiﬁed. Autotrophic CO2 ﬁxation pathways differ in their products and intermediates, and organisms belonging to different phylogenetic groups and living in different environments have different biosynthetic requirements. Traditionally, the costs of autotrophic carbon-ﬁxation mechanisms are judged by a comparison of how many ATP molecules are required to synthesize one molecule of triosephosphate from three molecules of inorganic carbon. However, the biosynthetic demands for triosephosphates and derived sugar phosphates differ dramatically. Plants using the CBB cycle synthesize huge amounts of sugars, whereas carbohydrates comprise just a minor fraction of archaeal cells. The direct energetic comparison of different cycles is sometimes questionable, as the reducing power of various electron donors is not the same. For example, the reduction potential of reduced ferredoxin is stronger than that of the reduced pyridine nucleotides. In growing cells, ferredoxin is often reduced by a hydroge-

3.3 Rules to Explain the Diversity

nase, but under a low hydrogen partial pressure the reduction of ferredoxin may be forced by energy-driven reverse electron transport from NAD(P)H, the costs of which are difﬁcult to evaluate [42]. This questions the possibility of comparing the energetic efﬁciency of, for example, the dicarboxylate/4-hydroxybutyrate cycle and the 3-hydroxypropionate/4-hydroxybutyrate cycle (see Table 3.1). The intrinsic properties of the CO2-ﬁxation-related enzymes may be important in terms of energetic costs of the whole pathway. A poor catalytic efﬁciency of RubisCO not only leads to a loss of energy, associated with photorespiration, but also pushes the organisms to synthesize special carbon-concentrating mechanisms (even C4 carbon ﬁxation and Crassulacean acid metabolism may be regarded as carbon-concentrating mechanisms). Moreover, carboxylases with a low catalytic efﬁciency need to be synthesized in large amounts; in the case of RubisCO, this may represent up to one-half of the soluble cellular protein. In other words, the synthesis of the carboxylase catalyst itself may devour a huge amount of energy. Not only the synthesis of the amino acids is costly; to build in one amino acid into protein, an additional four ATP are required. The amino acid demand for the synthesis of the additional enzymes of an autotrophic carbon-ﬁxation pathway may actually determine its energy costs. In terms of energy costs, the CBB cycle is one of the most expensive, as it requires nine ATP for the synthesis of one triosephosphate (see Table 3.1). However, most species that use this cycle are phototrophs (green plants, algae, cyanobacteria, purple bacteria); photosynthesis results in ATP production and, therefore, the bottleneck in the growth of phototrophs is normally not ATP supply, but rather the supply of phosphorus, nitrogen, or iron. Interestingly, the group of anoxygenic phototrophic bacteria adapted to low light intensities – the green sulfur bacteria [43, 44] – uses the reductive citric acid cycle for autotrophic CO2 ﬁxation. Indeed, the energy yield of their metabolism is much lower than that of the purple sulfur bacteria that use the CBB cycle. As both groups are successful in their environments, this simply reﬂects different ecological strategies applied by these organisms. Recent results reported by Markert et al. [45] are in agreement with this consideration; although, normally, an organism has only one pathway for autotrophic CO2 ﬁxation, an uncultured endosymbiont of a deep-sea tube worm uses both the reductive citric acid cycle and the CBB cycle, depending on the energy supply. In a high-energy situation the symbiont uses the CBB cycle, whereas under low-energy conditions it switches to the energetically more favorable reductive citric acid cycle. An important feature of the CBB cycle that may have ensured its success is its connection with carbohydrate metabolism. Carbohydrates are the most abundant biomolecules in plants, playing numerous roles as storage and structural compounds. The product of the CBB cycle, glyceraldehyde 3-phosphate, is a substrate of the gluconeogenetic pathway that directly connects CO2 ﬁxation with sugar synthesis; hence, this pathway may be advantageous for those organisms requiring a plentiful supply of carbohydrates. In four autotrophic CO2-ﬁxation cycles, succinyl-CoA plays a central role (Figure 3.7):

47

48

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

Acetyl-CoA

Glyoxylate

Oxaloacetate

Acetyl-CoA

HCO3CO2 HCO33-OH-Propionate

4-OH-Butyrate

Oxaloacetate 1

2

3

Citryl-CoA

5

4

CO2

HCO3Succinyl-CoA 1+4 1+3 2+5 2+4

Malyl-CoA

CO2

3-hydroxypropionate/malyl-CoA cycle 3-hydroxypropionate/4-hydroxybutyrate cycle reductive citric acid cycle dicarboxylate/4-hydroxybutyrate cycle

Figure 3.7 General scheme representing the strategy used by four different carbon dioxide-ﬁxation pathways [40]. These pathways have in common the formation of succinylCoA from acetyl-CoA and two inorganic carbons. The vertical arrows point to the CO2-ﬁxation products released from these

metabolic cycles. The combination of the metabolic modules 1 and 4 results in the 3-hydroxypropionate/malyl-CoA cycle; 1 and 3 = the 3-hydroxypropionate/4-hydroxybutyrate cycle; 2 and 5 = the reductive citric acid cycle; 2 and 3 = the dicarboxylate/4-hydroxybutyrate cycle.

i) In the 3-hydroxypropionate/malyl-CoA cycle, where acetyl-CoA/propionylCoA carboxylase(s) is (are) used as carboxylating enzyme(s) to form succinylCoA (route 1), from which acetyl-CoA is regenerated via malyl-CoA cleavage (route 4) and glyoxylate is the carbon-ﬁxation product (see Figure 3.4). ii) In the 3-hydroxypropionate/4-hydroxybutyrate cycle, where route 1 is used for succinyl-CoA formation, as in (i), but acetyl-CoA regeneration proceeds via 4-hydroxybutyrate (route 3), and acetyl-CoA is the product (see Figure 3.5). iii) In the dicarboxylate/4-hydroxybutyrate cycle, where acetyl-CoA is converted to succinyl-CoA using pyruvate synthase and PEP carboxylase (route 2); acetyl-CoA regeneration via 4-hydroxybutyrate is similar to the route in Sulfolobales (route 3) (see Figure 3.6). iv) In the reductive citric acid cycle, where succinyl-CoA is formed via route 2; however, succinyl-CoA is further reductively carboxylated to 2-oxoglutarate and isocitrate and converted to citrate, which is cleaved into acetyl-CoA and oxaloacetate (route 5) (see Figure 3.2). In principle, other combinations of the ﬁve partial routes indicated in Figure 3.7 are conceivable. The individual partial routes differ not only with respect to ATP

3.4 Evolutionary Aspects

requirement, but also with respect to oxygen sensitivity of the enzymes, and the use of reduced ferredoxin instead of NAD(P)H as reductant. Routes 2 and 5 represent typical “anaerobic” pathways that are unlikely to occur in strict aerobes, whereas routes 1, 2, and 4 are (micro)aerobic pathways. For example, the autotrophic Crenarchaea studied so far use one of two pathways, either the 3-hydroxypropionate/ 4-hydroxybutyrate (routes 1 + 3) or dicarboxylate/4-hydroxybutyrate (routes 2 + 3) cycle, depending on an aerobic or anaerobic mode of growth of the corresponding species. Although the reductive citric acid cycle (routes 2 + 5) functions in some bacteria requiring oxygen for growth (Aquiﬁcales), they require not only a reduced O2 content in the gas mixture, but also a high temperature (≥70 °C), further decreasing the oxygen concentration in the medium. Whether combinations others than those discussed exist remains unknown; however, interestingly, the combination of routes 2 and 4 (dicarboxylate/malyl-CoA cycle) was formerly proposed as an autotrophic CO2 ﬁxation pathway in Chloroﬂexus aurantiacus [46]. Autotrophic organisms preferentially incorporate 12CO2 (relative to 13CO2) into their biomass. This carbon isotopic fractionation cannot be achieved abiotically, and the occurrence of this effect in a specimen is an indication of its biological origin. Different CO2 ﬁxation pathways lead to the different isotopic depletions; for example, an isotopic signature of −20 to −30‰ is typical for the CBB cycle, of −2 to −13‰ – for the reductive citric acid cycle, the depletion greater than −30‰ is a characteristic of the reductive acetyl-CoA pathway, and values of −12.5 and −13.7‰ were reported for Chloroﬂexus aurantiacus in which the 3-hydroxypropionate/malyl-CoA cycle is operating [30, 47–50]. The 3-hydroxypropionate/4-hydroxybutyrate cycle and the dicarboxylate/4-hydroxybutyrate cycle both exhibited similar and very low fractionations, with reported values of between −0.2 and −3.6‰ [49]. Therefore, differences in the stable isotopic content may be used to suggest a speciﬁc mechanism of CO2 ﬁxation, and the pathways leave an isotopic signature in the geologic record of 13C-depleted sedimentary organic matter. Such results should be interpreted with great caution, however, because – as seen above – different pathways may have similar discrimination factors, and the discrimination may differ even among organisms using the same pathway. Furthermore, the contribution of different pathways leads to a mixed situation. For example, C4 and Crassulacean acid metabolism plants both have the CBB cycle, but their isotopic signatures are in the range of those for the reductive citric acid cycle. Moreover, a single amino acid substitution in the RubisCO active site can alter the discrimination signiﬁcantly, from −22 to −11‰ [50], thus questioning the possibility of identifying a particular pathway in the geological records by a fractionation value.

3.4 Evolutionary Aspects

The common ancestor of life was probably a chemoautotrophic hyperthermophilic anaerobe. This “metabolism ﬁrst” theory assumes that life started with catalytic metal sulﬁde surface/compartments in a hydrothermal-vent setting in the Hadean

49

50

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations

ocean [24, 51, 52]. According to this theory, inorganic carbon ﬁxation proceeded on minerals and was based on transition metal sulﬁde catalysts. Given the structural (and catalytic) similarity between the minerals themselves and the catalytic centers of the enzymes in the acetyl-CoA pathway, an attractive idea is that minerals catalyzed a primitive acetyl-CoA pathway [15]. In fact, experimental evidence supports this idea. For example, both the thioester acetyl methylsulﬁde and its hydrolysis product, acetate, can be produced from CO and CH3SH by using only Fe- and Ni-sulﬁdes as catalysts [23]. The following characters of the reductive acetyl-CoA pathway may indicate its antiquity [13]: the enzymes and electron carriers of this pathway are extremely oxygen-sensitive and can only function in strict anaerobes; the pathway occurs in both Bacteria and Archaea, and also in hyperthermophiles; it has minimal energy requirements; it allows the concomitant utilization of volatile reduced one-carbon compounds (e.g. CO, CH2O); metal sulﬁdes (in the form of different Fe–S centers) are important for the catalysis; it requires different ubiquitous, yet varied coenzymes, which probably preceded the more complex proteins as catalysts; and – most importantly- it combines energy metabolism and carbon assimilation mechanisms. The reductive acetyl-CoA pathway shares some of these features with the reductive citric acid cycle, which may also be regarded as primordial [14]. The dicarboxylate/4-hydroxybutyrate cycle also requires various iron–sulfur proteins such as ferredoxin, as well as thioesters to facilitate chemical reactions, and ﬁts well into a simple primordial carbonﬁxation scheme in an “iron–sulfur world” [24, 51]. At the same time, both phylogenetic analyses and general considerations denote the CBB cycle as a later innovation, the appearance of which might reﬂect the adaptation to oxygen [5, 13].

3.5 Chemical Aspects of CO2 Fixation

There are various aspects of autotrophic carbon ﬁxation linked to chemistry: 1) The need to reduce the accumulation of CO2 into the atmosphere requires new technologies capable of reducing the CO2 emission. One possibility would be to use CO2 as a raw material for syntheses, as Nature does on an incredible large scale using different mechanisms, as shown here. On the one hand, CO2 is available in virtually unlimited amounts, yet on the other hand it is very little reactive and its use often requires the input of energy. Hence, the use of CO2 in chemical syntheses requires dream reactions and molecular-deﬁned catalysts that are not yet available. This represents a great challenge for chemistry. 2) The use of CO2 in chemistry normally requires its interaction with metal centers of catalysts; one such example is the Kolbe–Schmitt carboxylation of phenol to produce salicylic acid. The potential of CO2 as a raw material in the synthesis of carboxylates, carbonates, or carbamates is rather limited. A future aim is the economically attractive synthesis of carboxylic acids, or optically

References

active β-hydroxycarbonic acids from ketones with CO2 and H2, under environment-friendly conditions. The development of catalysts for the synthesis of urea, formic acid, methanol, and cyclic carbonates are further future goals [53–55]. Unfortunately, the biological catalysts for similar reactions provide no direct models for chemical catalysts. 3) For various reasons, the carboxylases and reductive carboxylases of autotrophic CO2-ﬁxation pathways have not been applied in synthetic processes. Their high substrate speciﬁcity and target substrates make these biocatalysts unattractive, and therefore an alternative approach might be to develop existing autotrophic bacteria by metabolic engineering to produce useful chemical building blocks, and then to use their natural potential to synthesize an overwhelming variety of different compounds. The knowledge of new autotrophic pathways that involve completely different intermediates compared to the Calvin cycle opens new perspectives and potentials [7, 56]. 4) Another scientiﬁc “chemical problem of the century” is to mimic those reactions that, in chemoevolution, have resulted in the autocatalytic formation of various organic compounds from inorganic volcanic gases [53]. The existing autotrophic pathways may serve as guidelines showing in which direction to search for primordial processes.

Acknowledgments

These studies were supported by Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie.

References 1 Martin, W. and Schnarrenberger, C. (1997) Curr. Genet., 32, 1–18. 2 Bassham, J.A. and Calvin, M. (1957) The Path of Carbon in Photosynthesis, Prentice-Hall, Englewood Cliffs, NJ, p. 104. 3 Evans, M.C.W., Buchanan, B.B., and Arnon, D.I. (1966) Proc. Natl Acad. Sci. USA, 55, 928–934. 4 Amend, J.P. and Shock, E.L. (2001) FEMS Microbiol. Rev., 25, 175–243. 5 Tabita, F.R., Hanson, T.E., Li, H., Satagopan, S., Singh, J., and Chan, S. (2007) Microbiol. Mol. Biol. Rev., 71, 576–599. 6 Miura, A., Kameya, M., Arai, H., Ishii, M., and Igarashi, Y. (2008) J. Bacteriol., 190, 7170–7177.

7 Aoshima, M. (2007) Appl. Microbiol. Biotechnol., 75, 249–255. 8 Ellis, R.J. (1979) Trends Biochem. Sci., 4, 241–244. 9 Hanson, T.E. and Tabita, F.R. (2001) Proc. Natl Acad. Sci. USA, 98, 4397– 4402. 10 Finn, M.W. and Tabita, F.R. (2004) J. Bacteriol., 186, 6360–6366. 11 Sato, T., Atomi, H., and Imanaka, T. (2007) Science, 315, 1003–1006. 12 Mueller-Cajar, O. and Badger, M.R. (2007) BioEssays, 29, 722–724. 13 Fuchs, G.G. (1989) Biology of Autotrophic Bacteria (ed. H.G. Schlegel), Science Tech, Madison, Wisconsin, pp. 365–382. 14 Wächtershäuser, G. (1990) Proc. Natl Acad. Sci. USA, 87, 200–204.

51

52

3 Autotrophic Carbon Fixation in Biology: Pathways, Rules, and Speculations 15 Russell, M.J. and Martin, W. (2004) Trends Biochem. Sci., 29, 358–363. 16 Ivanovsky, R.N., Sintsov, N.V., and Kondratieva, E.N. (1980) Arch. Microbiol., 128, 239–241. 17 Schauder, R., Widdel, F., and Fuchs, G. (1987) Arch. Microbiol., 148, 218–225. 18 Hügler, M., Wisen, C.O., Fuchs, G., Taylor, C.D., and Sievert, S.M. (2005) J. Bacteriol., 187, 3020–3027. 19 Hügler, M., Huber, H., Molyneaux, S.J., Vetriani, C., and Sievert, S.M. (2007) Environ. Microbiol., 9, 81–92. 20 Strauss, G., Eisenreich, W., Bacher, A., and Fuchs, G. (1992) Eur. J. Biochem., 205, 853–866. 21 Ragsdale, S.W. and Wood, H.G. (1991) Crit. Rev. Biochem. Mol. Biol., 26, 261–300. 22 Thauer, R.K., Möller-Zinkhan, D., and Spormann, A.M. (1989) Annu. Rev. Microbiol., 43, 43–67. 23 Huber, C. and Wächtershäuser, G. (1997) Science, 276, 245–247. 24 Wächtershäuser, G. (2007) Chem. Biodivers., 4, 584–602. 25 Vorholt, J.A., Kunow, J., Stetter, K.O., and Thauer, R.K. (1995) Arch. Microbiol., 163, 112–118. 26 Schouten, S., Strous, M., Kuypers, M.M., Rijpstra, W.I., Baas, M., Schubert, C.J., Jetten, M.S., and Sinninghe Damsté, J.S., (2004) Appl. Environ. Microbiol., 70, 3785–3788. 27 Strauss, G. and Fuchs, G. (1993) Eur. J. Biochem., 215, 633–643. 28 Herter, S., Fuchs, G., Bacher, A., and Eisenreich, W. (2002) J. Biol. Chem., 277, 20277–20283. 29 Klatt, C.G., Bryant, D.A., and Ward, D.M. (2007) Environ. Microbiol., 9, 2067–2078. 30 Ivanovsky, R.N., Fal, Y.I., Berg, I.A., Ugolkova, N.V., Krasilnikova, E.N., Keppen, O.I., Zakharchuc, L.M., and Zyakun, A.M. (1999) Microbiology, 145, 1743–1748. 31 Alber, B.E. and Fuchs, G. (2002) J. Biol. Chem., 277, 12137–12143. 32 Hügler, M., Menendez, C., Schägger, H., and Fuchs, G. (2002) J. Bacteriol., 184, 2404–2410. 33 Friedmann, S., Alber, B.E., and Fuchs, G. (2007) J. Bacteriol., 189, 2906–2914.

34 Martins, B.M., Dobbek, H., Cinkaya, I., Buckel, W., and Messerschmidt, A. (2004) Proc. Natl Acad. Sci. USA, 101, 15645–15649. 35 Ishii, M., Miyake, T., Satoh, T., Sugiyama, H., Oshima, Y., Kodama, T., and Igarashi, Y. (1997) Arch. Microbiol., 166, 368–371. 36 Menendez, C., Bauer, Z., Huber, H., Gad’on, N., Stetter, K.O., and Fuchs, G. (1999) J. Bacteriol., 181, 1088–1098. 37 Berg, I.A., Kockelkorn, D., Buckel, W., and Fuchs, G. (2007) Science, 318, 1782–1786. 38 Hügler, M., Krieger, R.S., Jahn, M., and Fuchs, G. (2003) Eur. J. Biochem., 270, 736–744. 39 Tcherkez, G.G., Farquhar, G.D., and Andrews, T.J. (2006) Proc. Natl Acad. Sci. USA, 103, 7246–7251. 40 Huber, H., Gallenberger, M., Jahn, U., Eylert, E., Berg, I.A., Kockelkorn, D., Eisenreich, W., and Fuchs, G. (2008) Proc. Natl Acad. Sci. USA, 105, 7851–7856. 41 Valentine, D.L. (2007) Nat. Rev. Microbiol., 5, 316–323. 42 Hedderich, R. (2004) J. Bioenerg. Biomembr., 36, 65–75. 43 van Gemerden, H., and Mas, J. (1995) Anoxygenic Photosynthetic Bacteria (eds R.E. Blankenship, M.T. Madigan, and C.E. Bauer), Kluwer Academic Publishers, Dordrecht, The Netherlands. 44 Beatty, J.T., Overmann, J., Lince, M.T., Manske, A.K., Lang, A.S., Blankenship, R.E., Van Dover, C.L., Martinson, T.A., and Plumley, F.G. (2005) Proc. Natl Acad. Sci. USA, 102, 9306–9310. 45 Markert, S., Arndt, C., Felbeck, H., Becher, D., Siever, S.M., Hügler, M., Albrecht, D., Robidart, J., Bench, S., Feldman, R.A., Hecker, M., and Schweder, T. (2007) Science, 315, 247–250. 46 Ivanovsky, R.N., Krasilnikova, E.N., and Fal, Y.I. (1993) Arch. Microbiol., 159, 257–264. 47 Quandt, L., Gottschalk, G., Ziegler, H., and Stichler, W. (1977) FEMS Microbiol. Lett., 1, 125–128. 48 Holo, H. and Sirevåg, R. (1986) Arch. Microbiol., 145, 173–180. 49 House, C.H., Schopf, J.W., and Stetter, K.O. (2003) Org. Geochem., 34, 345–356.

References 50 McNevin, D.B., Badger, M.R., Whitney, S.M., von Caemmerer, S., Tcherkez, G.G.B., and Farquhar, G.D. (2007) J. Biol. Chem., 282, 36068–36076. 51 Wächtershäuser, G. (1988) Microbiol. Rev., 52, 452–484. 52 Martin, W., Baross, J., Kelley, D., and Russell, M.J. (2008) Nat. Microbiol. Rev., 6, 805–814.

53 Aresta, M. and Dibenedetto, A. (2002) Rev. Mol. Biotechnol., 90, 113–128. 54 Aresta, M. and Dibenedetto, A. (2007) Dalton Trans., 2975–2992. 55 Sakakura, T., Choi, J.C., and Yasuda, H. (2007) Chem. Rev., 107, 2365–2387. 56 Ishii, M., Chuakrut, S., Arai, H., and Igarashi, Y. (2004) Appl. Microbiol. Biotechnol., 64, 605–610.

53

55

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2 Joëlle Mascetti

4.1 Introduction

Despite the fact that carbon dioxide (CO2) is used in a great number of industrial applications, it remains a molecule of low reactivity, and methods have still to be identiﬁed for its activation. Both thermodynamic and kinetic problems are connected with the reactivity of CO2, and few reactions are thermodynamically feasible. A very promising approach to activation is offered by its coordination to transition metal complexes, as both stoichiometric reactions of C–C bond formation and catalytic reactions of CO2 are promoted by transition metal systems. Efforts to enhance the yield of hydrogen in water gas-shift (WGS) reactions have also been centered on CO2 interactions with transition metal catalysts. The coordination on metal centers lowers the activation energy required in further reactions with suitable reactants involving CO2, making it possible to convert this “inert” molecule into useful products. This chapter considers this important aspect of CO2 organometallic chemistry. As the subject has been reviewed previously, in 1995, 1996 and 1999 [1–5], attention has been focused here on the period between 1999 and early 2009. The ﬁrst section describes the general concepts of CO2 bonding to metals, and this is followed by details of electron-deﬁcient CO2–metal complexes characterized spectroscopically in low-temperature matrices, and stable complexes obtained at room temperature, in order to demonstrate the different coordination modes identiﬁed, trends along the Periodic Table, and theoretical contributions to the understanding of bonding in these systems. The subsequent sections involve the reactivity of coordinated CO2, focusing initially on reactions proceeding via O transfer or electrophilic or nucleophilic attacks on a bent coordinated OCO moiety with increased C–O distance (i.e., “true” reactions of coordinated CO2). Examples are then provided of oxidative coupling, reduction reactions, catalytic processes, and bioinspired reactions where the formation of an intermediate metal CO2 complex is suspected in the reaction mechanism.

Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

56

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

4.2 Carbon Dioxide Bonding to Metals

Carbon dioxide is a 16e− molecule that belongs to the D∞h symmetry group and is linear in its ground state. It is a nonpolar molecule containing two polar C=O bonds, with two sets of orthogonal π orbitals. Initially thought to be a poor ligand, it is now evident that CO2 exhibits several coordination sites and a great variety of coordination modes in its complexes. The carbon atom (LUMO orbitals) has a Lewis acid character and can be described as an electrophilic center, whereas the oxygens (HOMO orbitals) are weak Lewis bases and described as nucleophilic centers. It is worth noting that most of CO2 catalytic reactions require a simultaneous acid–base activation, with the carbon atom and one of the oxygen atoms involved in the interaction with the metal. The two double C=O bonds contain π electrons that can interact with d electrons of transition metals in a Dewar–Chatt– Duncanson bonding scheme. When the LUMO orbitals of CO2 are occupied (via electron transfer), the lowest energy state corresponds to a bent geometry. As an example, the radical anion CO2−• is a bent molecule with an equilibrium angle of 134 °. So, any interaction of carbon dioxide with a metal will induce a loss of linearity. There are four basic modes of CO2 coordination, as described in Figure 4.1. For the η1-C coordination mode, there is a strong charge transfer between a d2z metal orbital and the anti-bonding π* orbital of CO2. This bonding mode is preferred with electron-rich metals, and may be facilitated by an additional weak interaction between one or two oxygen atoms of CO2 with a Lewis acid center located in the coordination sphere of the metal. In the η2(C,O) bonding mode, there is a double bonding scheme with a σ bond from the π orbital of CO2 to an empty d2z metal orbital, together with a “back-bonding” from a ﬁlled dxy metal orbital to the empty π* CO2 orbital. The η1(O) end-on coordination mode is preferred with electronpoor metals, and the CO2 molecule can remain linear or be weakly bent. The η2(O,O) coordination mode can be described as a metal carboxylate with an ionic

O O

C

O

O

O

C

Lewis acid (at C)

C

O

M

η1O

η1C M

LUMO

+

M

M

O

O C

O

C

O

Lewis base (at O)

O C

η2CO HOMO

Figure 4.1

O

M

CO2 coordination to a single metal center.

η2O,O

4.2 Carbon Dioxide Bonding to Metals

bond M+ CO2− and is often encountered with alkali or alkaline-earth metals or, in the case of CO2 adsorption, with metal surfaces. The interaction of CO2 with transition metal complexes has been the subject of extensive studies [1–6], both experimental and theoretical. It is remarkable that only a few authenticated examples of metal complexes of CO2 have been reported in the literature. The ﬁrst side-on complex to be structurally characterized was reported by Aresta and colleagues [7, 8] with a Ni atom coordinated to one CO2 and two phosphine ligands in an almost planar environment: (PCy3)2Ni(CO2). Coordinated CO2 has two nonequivalent CO bonds (1.17 and 1.22 Å) and an OCO angle of 133 °. Later on, an iron complex with a trigonal bipyramidal geometry and CO2 lying in the equatorial plane was synthesized by Karsch [9]: (PMe3)4Fe(CO2). In this complex, CO2 has two almost equivalent relatively long CO bonds (1.25 and 1.28 Å) and a small OCO angle of 124 °, very close to that of a CO2− moiety. The ﬁrst Ccoordinated complex to be structurally characterized was reported by Herskovitz et al. [10] for Rh: Rh(diars)2Cl(CO2), with CO bonds lengths of 1.20 and 1.25 Å and an OCO angle of 126 °. Other CO2 complexes of Nb [11], Mo [12, 13], and Co [14] were structurally characterized during the 1980s and, more recently, a CO2 complex of Ru [15] has been described. The CO2-bridged polymetallic complexes involve coordination of the carboxyl carbon to one metal and bonding of one or two carboxyl oxygens to a second (or third) metal center, leading to compounds of the μm-ηn type, as described in Figure 4.2. The ﬁrst bridged CO2 complexes to be structurally characterized were Fe–Re, Ir–Zr, and Rh–Os bimetallic complexes, or polymeric Co, Os, and Ru clusters (for a review, see Ref. [4] and references therein for details). When X-radiography studies are not possible, theoretical calculations and spectroscopic studies can help in the structural characterization of CO2 complexes. The most largely used as a diagnostic tool is infrared (IR) spectroscopy, both for the quantitative determination of CO2 and for structural analysis. Because CO2 has a center of symmetry, the asymmetric stretching (2340 cm−1) and the degenerate bending modes (667 cm−1) are IR-active and Raman-inactive. The C=O symmetric stretching mode is IR-inactive, whereas a Fermi resonance is observed as a doublet located at 1285 and 1388 cm−1 in the Raman spectrum of CO2. When the LUMO orbitals are occupied (as may occur in the radical anion CO2−•, in CO2 adducts, or in electronically excited CO2) it induces the bending of the molecule and an increase in the CO bond length, and, consequently, large modiﬁcations in the IR spectrum of the CO2 moiety. The anti-symmetric νa(C=O) stretching mode is lowered in the region 1500–2000 cm−1, the symmetric νs(CO) stretching mode becomes IR-active and observed in the region 1400–1100 cm−1, the bending mode δ(OCO) is shifted from 667 cm−1, and additional modes (metal– C and/or metal–O stretching modes, C=O out-of-plane bending mode) may appear in the low-frequency region, between 300 and 800 cm−1. The positions and intensities of these modes can provide some information on the bonding mode of CO2

57

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

58

binuclear complexes

M O

O O

M

C

O

O C

M

O C

M

μ2−η2C,O

M

O

O M

μ3−η3

Figure 4.2

O,O'

O

M O

O M

C

M

2

tetranuclear complexes M

C

M

μ2−η

M

trinuclear complexes

O

O

μ2−η4

M

μ2−η3

M

C

M

M

O

M

μ3−η4

M

C

μ4−η5

Coordination modes in polynuclear metal–CO2 complexes.

in the complex, together with an estimation of the OCO angle value. Together with the use of isotope-labeled CO2, Fourier transform infrared (FTIR) spectroscopy can then be very precise and helpful in the characterization of CO2–metal complexes. Jegat et al. [16] have shown that, if ΣΔν represents the sum of the frequency shifts observed on the two ν(CO) stretching modes in isotope-labeled 13C and 18O complexes, then the following relationships are to be expected:

• • •

side-on coordination: ΣΔν(13C) > ΣΔν(18O) and ΣΔν(18O) < 60 cm−1 η1-C coordination: ΣΔν(13C) > ΣΔν(18O) with 60 < ΣΔν(18O) < 70 cm−1 end-on coordination: ΣΔν(13C) < ΣΔν(18O) with ΣΔν(18O) > 70 cm−1

In C-coordinated complexes, the frequency splitting between the two ν(CO) stretching modes is less than 400 cm−1, and in end-on complexes, the out-of-plane bending mode γ(C=O), located in the 500–650 cm−1 region, exhibits a preponderant 18 O effect, in contrast with other coordination modes, where a larger 13C effect is observed (namely Δγ(13C) = 10 to 20 cm−1, versus Δγ(18O) = 5 cm−1). Nuclear magnetic resonance (NMR) spectroscopy is also largely used to characterize CO2 complexes. The 13C NMR spectrum of CO2 dissolved in a nonpolar solvent shows a resonance at 124 ppm, which is shifted when CO2 is bonded to a metal center. Depending on the mode of bonding, the shift may be up or down ﬁeld, and may vary from a few ppm up to several hundreds of ppm. A few examples are given below for different types of bonding.

4.3 Synthesis and Structure of CO2 Complexes

4.3 Synthesis and Structure of CO2 Complexes

Here, the different identiﬁed coordination modes of CO2 with transition and nontransition metals are described, together with trends along the Periodic Table, and theoretical contributions to the understanding of bonding in these systems through three types of study: (i) low-temperature matrix isolation spectroscopy of electrondeﬁcient metal/CO2 moieties; (ii) theoretical studies of reactions of metals with CO2; and (iii) the synthesis of stable complexes. No information is provided on gas-phase studies, nor on the surface chemistry of CO2 (such details are provided in Refs [4, 17]). 4.3.1 Low-Temperature Matrix Isolation and Theoretical Studies

Matrix isolation is a technique whereby an unstable or reactive chemical species can be immobilized by cocondensing it at very low temperature with copious quantities of an inert substance, such as a noble gas [18]. A large number of matrix isolation studies have focused on the reactions of atoms, which can be thermally produced without any excess kinetic energy. Due to the combination of the matrix “cage effect” and low temperature, dynamic processes are then often hindered, and even a small barrier to reaction appears as a mountain at 10 K when kT = 0.08 kJ mol−1. However, this provides access to loosely bound complexes, which may act as precursors to insertion or other reactions. The technique of laser ablation is now widely used in many of these studies to generate metal atoms, and also to initiate atomic reactions in cryogenic matrices. In this case, it must be borne in mind that atoms have an excess kinetic energy that may allow them to overpass energy barriers and prevent the observation of reaction intermediates, such as coordination complexes. Photochemical effects may also occur, due to the light being emitted during the ablation process. Infrared spectroscopy remains, together with theoretical [mainly density functional theory (DFT)] calculations, the workhorse of these studies, allowing reasonably certain identiﬁcation. Therefore, a convenient route to the preparation of metal–CO2 complexes and their identiﬁcation by means of spectroscopy might be achieved through the condensation reaction of metal atoms with CO2 at cryogenic temperatures. As this particular aspect of metal–CO2 complex formation has not been reviewed speciﬁcally, the following section will include a more important period than 1999–2009, though not exhaustively. The ﬁrst matrix isolation spectroscopic study conducted by M.E. Jacox et al. [19] on alkali atoms and CO2 co-deposited at 14 K in Ar matrices, showed IR absorptions near 1600 cm−1 that were assigned to the anti-symmetric stretching mode ν3(OCO) of an M+ …CO2− ion pair, with an OCO valence angle near 130 ° and a signiﬁcant contribution of molecular aggregates to the observed spectrum. However, the ﬁrst attempts to prepare isolated metal–CO2 complexes in a cryogenic matrix appeared during the early 1980s. An example was that of Kafaﬁ et al. [20], who condensed alkali atoms (Na, K, Cs) with pure CO2, CO2/Ar or CO2/N2

59

60

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

mixtures and formed M+ CO2− (with a η2-O,O coordination and C2v symmetry) and 2− M2+ 2 CO2 (with C2v and Cs structures). Shortly before this, Hauge et al. had claimed that CO2 was reductively coupled by Li to give LiC2O4 by condensing Li atoms in CO2/Ar mixtures [21]. Later on, Manceron studied very carefully the reactions of Li atoms with CO2 in Kr [22], and observed the formation of two isomers of LiCO2 with η1-O coordinated bent structures, namely LiC2O4 and Li2C2O4. The η1-O coordinated Li–OCO complex has long been a unique example of end-on coordination, and the question arose of whether a true oxalate (with a C–C bond formation) was formed, or not. The result of these studies was that a quite rich reactivity was observed (except for Na), with the formation of M–CO2 complexes with various structures (C2v, Cs), leading to the formation of oxalates MC2O4 by a concentration effect or annealing, and to the decomposition in carbonates M2CO3 and CO by annealing above 200 K. Surprisingly, sodium atoms appeared almost nonreactive, except in neat CO2 matrices. The decreasing order of reactivity of alkaline metals with CO2 was Li > Cs > K > Na. Other metals, such as alkaline-earth Be [23], Mg [24–26], Ca [27], or Group 13 metals such as B [28, 29] and Al [30], have been studied, both experimentally and/ or theoretically. The reaction M + CO2 → MO + CO is calculated for all laser ablated alkaline-earth atoms as endothermic by 26, 66, and 35 kcal mol−1 for Be, Mg, and Ca, respectively [23, 25, 27]. The reactions proceed through the formation of cyclic (η2-O,O) MOCO structures via barriers of 23, 20, and 14 kcal mol−1, respectively. A previous experimental study of Mg + CO2, conducted with thermal Mg atoms [24], led to the formation of a MgOCO C2v cyclic structure with a calculated OCO angle of 128 °. The IR bands of the complex were observed at 1580, 1385, and 866 cm−1, although it should be noted that, in line with calculations and as previously observed for Na atoms, only a small amount of the Mg atoms was reactive, even in neat CO2. Interestingly, the condensation of Mg atoms with CO2/C2H4/Ar mixtures exhibited IR bands at 1768, 1284, 1256 cm−1, that were assigned to the formation of a ﬁve-membered cycle MgC2H4CO2 through the formation of a C–C bond between ethylene and CO2 (binding energy: 18 kcal mol−1). Similar attempts to obtain a reductive coupling between ethylene and CO2 with Ni atoms in cryogenic matrices were unsuccessful. The reforming of CO2 to CO is signiﬁcantly enhanced in the presence of B [28, 29], as the reaction B + CO2 → BO + CO becomes exothermic by 64 kcal mol−1. The recombination of BO and CO, leading to OBCO, is predicted to be barrier-free [29], and the OBCO (νCO = 1863 cm−1) product detected in matrix isolation experiments [28] was probably a secondary product of the BO + CO recombination. The codeposition of Al atoms with CO2 in Ar matrices [30] leads to the formation of AlCO2 molecules, which interconvert into two isomers: the low-temperature isomer presents a Cs symmetry, with two nonequivalent CO bonds, whereas the highertemperature (25 K) isomer has a ring structure in which the metal interacts with both oxygen atoms. For Al clusters, a reductive elimination is observed upon warming up above 30 K, yielding Al2O and CO. When investigating transition metals, Ozin et al. made the ﬁrst attempts with Au and Ag atoms [31], and observed very labile MCO2 π complexes by means of

4.3 Synthesis and Structure of CO2 Complexes

Raman and UV-visible spectroscopy, but no precise characterization was made. A report was made in 1981 where the IR spectrum of Cu atoms deposited with CO2 at 80 K was interpreted in terms of the formation of a π-coordinated complex between CO2 and zerovalent copper [32]. Almond et al. [33] prepared a (CO2) M(CO)5 molecule (M = Cr, W), that led to the formation of CO and oxometal carbonyl under UV irradiation. The ﬁrst complete study of the reactivity of CO2 with the ﬁrst row of transition metals was made by Mascetti et al. [34, 35]. Here, it was shown that the late transition metal atoms (Fe, Co, Ni, and Cu) formed one-to-one M(CO2) complexes, where CO2 was bonded in a side-on (Ni), end-on (Cu), or C-coordinated (Fe, Co) manner, while the earlier metal atoms (Ti, V, and Cr) spontaneously inserted into a CO bond to yield oxocarbonyl species OM(CO) or OM(CO)(CO2). Normal coordinate analysis showed that the force constants of CO bonds were signiﬁcantly decreased by 50%, compared to free CO2, and that the OCO angle was bent between 120 and 150 °. Since then, most studies have associated experimental matrix isolation studies with DFT calculations: Sc [36–38], Ti [36, 39–41], V [40, 42], Cr [43], Mn [43], Fe [43, 44], Co [43, 45], Ni [36, 43, 46–48], and Cu [36, 43, 49] for the ﬁrst-row transition metals. This has allowed for the prediction of reaction mechanisms, a better description of the bonding modes, and a more reliable identiﬁcation of the species observed in experiments. Examples of this include titanium, nickel, and copper. The interaction of s2d2 and s1d3 Ti atoms with CO2 has been studied by Papai et al. [39] using DFT calculations. It has been shown that the ground-state Ti atom could insert with no energy barrier into a CO bond, and this would result in an OTiCO insertion product (see Figure 4.3). The comparison of the calculated data with those from previous matrix isolation experiments [35, 40, 50] revealed that the insertion product formed in a low-temperature argon matrix with thermal- [35] or laser-evaporated [40] Ti atoms corresponded to the singlet state OTiCO. Ti(CO2) complexes, in various coordination modes, were located on the triplet and quintet potential surfaces, from which the triplet state (O,O) coordination mode was shown to be the most stable, but this lay above the OTiCO molecule by about 30 kcal mol−1. Later, Mebel et al. [41] conducted a very detailed study of the same reactions, and reached the conclusion that the most energetically favorable reaction mechanism was an insertion of the Ti atom into a CO bond via a η2(C,O) -coordinated Ti(CO2) complex, to produce the triplet OTi(CO) molecule. A comparison of the reaction mechanisms for alkaline earth and early transition metal atoms indicated that, although all of them could enhance CO2 reforming into CO, the early transition metal atoms (Sc, Ti, V) were the best for this purpose. The case of nickel is completely different. Although laser-ablated atoms exhibit insertion reactions [43], thermal atoms are almost nonreactive [35] unless working in a neat CO2 matrix [35, 48] or using a coreactant such as N2 [51]. Andrews et al. [43] have shown that laser-ablated late transition metals (from Cr to Ni), react with CO2 to give the insertion products OMCO, as observed in solid argon matrices. Mebel et al. [47] showed, via DFT studies, that this reaction occurred preferentially in the triplet electronic state, through the formation of a cyclic four-membered

61

62

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

Figure 4.3 Potential energy curve for the Ti + CO2 → OTiCO insertion reaction based on DFT calculations.

ring NiOOC intermediate, with an energy barrier of about 23 kcal mol−1, although the basis sets used did not allow accurate results to be obtained. Later on, Hannachi et al. [48] revisited the theoretical study of this system by using combined B3LYP/CCSD(T) calculations, and showed that the insertion–elimination route Ni + CO2 → NiO (3Σ−) + CO (1Σ+) was the most favorable mechanism (Figure 4.4). In this case, the metal insertion reaction is initiated by an electron transfer from the Ni atom to the CO2 molecule, after which the insertion and oxygen abstraction steps take place in a concerted fashion, along with the charge-transfer processes. So, it is clear that for nickel, the insertion reaction is not preceded by the formation of a η2-CO2 complex, as with Sc, Ti, or V. Likewise, in contrast to the Sc, Ti and V reactions, the insertion of Ni into CO2 is an endothermic reaction (by 15.4 kcal mol−1) with respect to Ni (s1d9) + CO2 and, as a consequence, was not observed for thermally evaporated nickel atoms, which have insufﬁcient kinetic energy to overcome the barrier of reaction. Similar to nickel, thermal copper atoms are not especially reactive towards CO2: the formation of a weak 1 : 1 end-on complex has been observed in neat CO2 matrix by Mascetti et al. [35], characterized by an important charge transfer from copper to CO2 [52]. Later, theoretical investigations were conducted by Hannachi et al. [49] which showed that the insertion of Cu into the CO bond of CO2 was direct. with an activation energy of 61 kcal mol−1 and also endothermic, by 30.6 kcal mol−1. Although no Cu(CO2) complex was found, a weak η1O -Cu2 (OCO) was calculated, which suggested that the species observed experimentally in the matrices was in

4.3 Synthesis and Structure of CO2 Complexes 60.0

Ni

59.9

3A" surface

2.111

Relative energy (kcal/mol)

3A' surface

NiO(3Π)+CO

1A' surface

51.7

45.0 40.0

20.0

O

14.3 = == Ni+CO2 0.0

?

h2oo (3B2)

O

1.846

1.700

Ni 179.9

h2co (1A')

1.129 O

1.838

1.131 C 180.0

O

ONiCO (3A') (3A")

Ni 1.864

-7.8

C 170.7

ONiCO (3A")

15.4 ONiCO

h2oo(3B2)

1.2287

145.0

(3A')

17.4

17.2

Ni

1.665

37.4 ONiCO

O

C

NiO(3Σ-)+CO

34.6

132.1

O

h2co (1A')

O

1.249

1.904 C 147.6

1.180 O

Figure 4.4 Energy diagram of the Ni + CO2 → NiO + CO reaction based on DFT calculations.

fact a Cu2 end-on complex [53]. The formation of OCuCO− anions was observed with laser-ablated Cu atoms by Andrews et al. [43], produced by the addition of Cu to CO2− , formed by electron capture during the experiments. Some metals from the second and third rows of the Periodic Table, that are largely used in catalytic systems and reduction processes, have also been studied; these include Zr [54], Ru [55], Rh [45], Pd [56], Nb [57], Ta [58], Re [59], and Os [55]. For laser-ablated Zr, Nb, and Ta atoms, Qin et al. have observed the formation of OM(CO) and O2M(CO)2 insertion products. Subsequent calculations proved that the insertion process was barrierless for Zr, whereas Nb and Ta atoms insertions proceeded via electron transfer from the metal to CO2 and the formation of adduct complexes as intermediates. Laser-ablated Co and Rh atoms, when studied by Xu et al., showed similar reactions with the formation of the following neutral species: OM(CO), O2M(CO), and OCo2(CO). For laser-ablated Ru and Os atoms cocondensed with CO2 in Ar and Ne matrices, Andrews et al. have identiﬁed several neutral insertion species, including OM(CO), O2M(CO), and O2Os(CO)2, formed by the insertion of OOs(CO) into a second CO2 molecule, and OCRu(O2)(CO) by the addition of a CO2 molecule to ORu(CO). Although, osmium has been shown to be more reactive than ruthenium, both metals have the potential to be good catalysts for CO2 reduction. Similar results have been obtained for Re by the same group. A theoretical study performed by Liu et al. [56], of CO2 coordination in a series of phosphine-substituted Pd complexes, showed that in all cases the η2C,O side-on bonding mode was the lowest in energy, independently of the basicity

63

64

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

or the bulkiness of the ligand phosphine; however, the relative stabilities were different. Several studies have also been performed on heavy metals such as lanthanides or actinides. In the case of La atoms [60], the formation of OLa(CO) as the primary product was observed, whilst UV-visible rearrangements have been seen to lead to the formation of La(η2-CO)O and OLa(η2-CO). Finally, a comparison of the decreasing CO stretching frequency of OMCO in the series Sc, Y, La, indicated an increase in the metal d orbital → CO π* back-donation, going from Sc to La. In the case of uranium [61], Andrews et al. showed that pulsed laser-evaporated U atoms were sufﬁciently energetic to insert into CO2 upon condensation in argon matrices, forming OUCO (1799, 804 cm−1), and with a larger charge transfer from OU to CO than for OBCO [29]. Further reaction with CO molecules isolated in the matrix led to the formation of OCU(O)CO species. It appears, then, that the most-often observed reactions of a metal atom with CO2 in a low-temperature matrix are the formation of a metal complex, and/or the insertion into one CO bond of CO2. These studies, which were conducted at low temperature on “naked” metal atoms, could not reproduce the reactions obtained with “real” metal complexes containing ligands, which can in turn inﬂuence further reactions with CO2 (some examples are provided in Sections 4.3 and 4.4). However, with the assistance of theoretical calculations, the studies have allowed the identiﬁcation of general trends in the Periodic Table, as well as a description of the different CO2 bonding modes through the vibrational analysis of isolated M(CO2) moieties. 4.3.2 Synthesis of Stable Complexes

The synthesis of stable CO2 complexes usually proceeds via four different methods: (i) the substitution of labile ligands (as for Ni(PCy3)2(CO2) [7]); (ii) addition to d16 complexes (as for Rh(diars)2Cl(CO2) [10]); (iii) insertion into binuclear complexes (as for Co(Pr-salen)K(CO2) [14]) and the reduction of complexes (as for Nb(CO2) (C5H4Me)(CH2SiMe3) [11]); or (iv) by the in situ synthesis of CO2 as for ((CO)5Re2 (CO2)2Re(CO)4)2 [62]). The most important of these reactions, however, are substitutions and additions, as these also play a decisive role in the catalytic and stoichiometric reactions of CO2. At this point, only those compounds that have been clearly identiﬁed as having CO2 bound to metal centers through one at least of its atoms will be considered. Some relatively recent reviews by Gibson [2] and Leitner [3] in 1996, and by Gibson [4] and Yin [5] in 1999, described the most recent developments of the organometallic chemistry of CO2 up until 1998; hence, these will not be discussed here in detail, other than to mention any relevant aspects of CO2 activation. Rather, attention will be focused on data produced between 1999 and 2009. Perhaps the most important report made during the past decade has been the ﬁrst synthesis of stable end-on CO2 coordinated complexes. The ﬁrst such complex was a linear η1-O-coordinated CO2 bound to uranium, obtained by

4.3 Synthesis and Structure of CO2 Complexes

Castro-Rodriguez et al. in 2004 [63]. Subsequently, a linear μ(O, O′)-coordinated CO2 to magnesium was synthesized by Chang et al. in 2005 [64] whilst, very recently, in 2008, Gao et al. reported a μ-η1,η1-OCO linear coordination mode of CO2 in inorganic polyoxoanions in water [65]. 4.3.2.1 End-On Complexes The electron-rich six-coordinate (AdArO)3tacnUIII complex (where (AdArOH)3tacn = 1,4,7-tris(3-adamantyl-5-tert-butyl-2-hydroxybenzyl)1,4,7-triazacyclononane) was shown to react with CO2 to yield (AdArO)3tacnUIV(CO2), a colorless complex in which CO2 was linearly coordinated to U through one oxygen [63]. When this complex was characterized with X-ray diffraction (XRD), the OCO angle was close to linear (178 °), the U–OCO group had a U–O bond length of 2.351 Å, a short C–O bond length of 1.122 Å, and a longer terminal CO bond length of 1.277 Å. A bonding model UIV=O=C•–O− ↔ UIV-O≡C–O−, where the uranium center is oxidized to UIV and the CO2 ligand reduced by one electron, was supported by electronic and vibrational spectroscopic studies. Both, visible and near-infrared (NIR) band intensities and positions were characteristic of the UIV f2 ion, while the IR spectra exhibited a ν(CO) band at 2188 cm−1, which was shifted by 60 cm−1 when using 13CO2. In 2005, Chang et al. [64] reported a one-pot synthetic pathway to novel aluminum–magnesium complexes containing CO2 in the linear μ(O,O′) bonding mode. For this, one equivalent of AlR3 (R=Me, Et) was added to a solution of Mg(N(SiMe3)2)2 in tetrahydrofuran (THF) at room temperature, and an excess of CO2 was bubbled through the stirred, ice-cooled mixture to yield compounds 1 (R = Me) and 2 (R = Et): (R2Al(μ-NSiMe3)(μ-OSiMe3)Mg(thf)2(μ-O2C))3. The proposed reaction pathway suggested a ﬁrst reaction between CO2 and Mg(N(SiMe3)2)2 to give an oxo-transfer product Mg(N(SiMe3)2)(OSiMe3), which was assumed to form a bridged Al–Mg intermediate with AlR3; this then lost a ligand from the Mg center and was attacked by a second CO2 molecule, with the oxygen atom as a weak electron donor. Finally, the CO2 acted as a bridging ligand to form a trimer (Figure 4.5). The structures of 1 and 2 were conﬁrmed by XRD, with a C3 symmetry axis, and three almost linear CO2 bridges that formed a 12-membered ring. The CO bond lengths were very close to that of free CO2 (between 1.149 and 1.233 Å), which suggested the presence of double C=O bonds. The OCO angles (between 169.6 and 175.3 °) were close to 180 ° and suggested sp-hybridized C atoms. The Mg←O=C=O→Mg moieties showed some disorder, with a shape like an hourglass. The 13C NMR spectra of 1 and 2 showed peaks that were slightly shifted (ca. 121 ppm) relative to free CO2 (124 ppm), whereas the side-on transition metal complexes usually showed 13C resonances at around 200 ppm. The IR spectra exhibited strong absorptions at 2267 and 2275 cm−1, respectively, these being slightly red-shifted by 73 and 65 cm−1 relative to free CO2, whereas the side-on complexes usually exhibited shifts greater than 300 cm−1. Only a small amount of electron density was transferred from the coordinated CO2 to the more electropositive Mg atoms in molecules 1 and 2 (and less than that observed in the uranium complex described above).

65

66

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

Figure 4.5

Schematic representation of Al–Mg complex with linearly bridging CO2 [64].

By bubbling CO2 through an aqueous solution containing (H2PMo11CoO40)5−, Gao et al. were able to obtain the CO2 adduct compound (C3H5N2)3(C3H4N2) (PMo11CoO38(CO2))·4 H2O, where single crystal XRD showed that the CO2 was μη1,η1-coordinated [65]. The skeleton was composed of (PMo11CoO38)n polymeric chains, bridged by the CO2 ligand in an axial direction. The CO2 groups showed slightly bent arrangements, with OCO angles of 158.7 ° and a CO bond length of 1.287 Å. The two O atoms of CO2 linked MoVI and CoII from two neighboring polyoxoanions, respectively. The IR spectrum exhibited an absorption band at 2169 cm−1, shifted to 2108 cm−1 by 13C labeling. In the NMR 13C spectrum, the chemical shift for the C atom of CO2 was 113.86 ppm, appearing in an up-ﬁeld shift relative to free CO2. Similar results have been obtained for the analogue compound (C3H5N2)4(SiMo11CoO38(CO2))·4 H2O. Beyond the fact that it was one of the few examples of stable end-on CO2 complexes, this synthesis provided a promising route to capture CO2 by simple inorganic polyoxoanions in an aqueous medium. Interestingly, in 1998 Kozik et al. [66] had described quite different results on the interaction of CO2 with various tungsten heteropolyanions. Reactions with CO2 occurred for germano- and silicotungstates when substituted with Co(II), Ni(II) and Mn(II). Although no structural data were available, multiple spectroscopic evidence indicated that reversible CO2 complexes had been formed, with a direct η1-C metal– carbon bond (the so-called C-coordination), and with the O atoms of CO2 being involved in H bonds with water or alcohol included in the structure. The IR spectrum of the tetraheptylammonium salt of α-(SiW11O39Co(CO2))6− showed strong bands located around 1675 and 1350 cm−1, respectively shifted to 1633 and 1340 cm−1 with 13CO2, and to 1657 and 1331 cm−1 with C18O2. The 13C NMR study was the ﬁrst to report the existence of octahedral paramagnetic Co(II) complexes, with two considerably downﬁeld shifted signals observed at 792 and 596 ppm, due to two different types of H bonding, and with sensitivity to the temperature.

4.3 Synthesis and Structure of CO2 Complexes

4.3.2.2 Side-On Complexes A new bis(carbon dioxide) side-on adduct of molybdenum: transMo(CO2)2HN(CH2CH2PMe2)2(PMe3) has been synthesized from transMo(CO2)2(PMe3)4 [12] and structurally characterized by Carmona et al. [67]. Similar to the parent complex, this new compound had a distorted octahedral geometry with two staggered CO2 ligands that were strongly bonded to Mo through one π-CO bond, in a side-on manner. The CO2 groups had CO bond lengths ranging from 1.199 to 1.284 Å, and OCO angles equal to 131 °, whereas the Mo–C distance was 2.08 Å. The IR data showed red-shifted bands at 1660, 1155, and 1100 cm−1. The 13C NMR spectra revealed a ﬂuxionality that consisted of a synchronous motion of the two CO2 ligands, rotating in the same direction, with two signals centered at 213 and 217 ppm. One of the reactions of coordinated CO2 is its reduction to CO by means of oxygen atom transfer to another substrate (an oxophilic metal or a readily oxidized ligand). However, CO2 can also act as its own oxygen sink, giving rise to CO + CO23 − . The carbonyl carbonate compound trans-Mo(CO)(CO3)HN(CH2CH2PMe2)2(PMe3) has also been prepared from the tetrakis phosphine carbonyl carbonate parent compound. In addition, the reactivity of Mo(CO3)(CO)(PMe3)4 towards CO has been studied: a dicarbonyl complex Mo(CO3)(CO)2(PMe3)3 was formed at room temperature. Interestingly, this latter compound could undergo a conproportionation reaction and liberate CO2 upon heating under a carbon monoxide atmosphere; this was seen as a rare example of the reversibility of the disproportionation process. 4.3.2.3 Bridged Complexes Sadighi et al. reported in 2007 [68] the reduction of CO2 by a (N-heterocyclic carbene)Ni(0) complex that gave rise to a new μ-η2, η2-CO2 coordination geometry at a dinickel core: ((IPr)Ni)2(μ-CO)(μ-η2,η2-CO2), with IPr = 1,3-bis(2,6diisopropylphenyl)imidazol-2-ylidene. The reduction of CO2 was evident from the presence of the bridging CO in the complex, which showed two NMR 13C signals at 172.6 ppm (CO2) and 246.4 ppm (CO). This complex also exhibited strong IR absorptions at 1773, 1630, and 1205 cm−1, shifted to 1731, 1586, and 1183 cm−1 upon 13C isotopic substitution. The most remarkable feature of the crystal structure was one CO2 ligand bridging two nickel atoms via its two π-CO bonds in a side-on manner (Figure 4.6). The bridged CO2 was bent at 133.4 ° and the CO bond lengths were 1.255 and 1.257 Å, which were much longer than for free CO2 (1.16 Å) and for the bridged CO ligand (1.184 Å). Other Ni(0) complexes, such as biscarbene (IMe)2Ni(0) formed analogous complexes (IMe = 1,3-dimesitylimidazol-2-ylidene). As with the previous example, a carbonyl carbonate Ni(II) isomer was formed by a disproportionation reaction during the synthesis. 4.3.2.4 Bridged Complexes Obtained by In-situ Synthesis A few stable CO2 complexes resulting from a formal oxidation of metal carbonyl complex have been described. Although not strictly true CO2 complexes, these are important intermediates in the photocatalytic and electrocatalytic reductions of CO2 to formate and CO, and also in biologic systems. A few examples of such species are described below.

67

68

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

O(1)

C(6)

C(3) Ni(1)

C(1)

Ni(2)

C(2) O(3)

Figure 4.6

O(2)

The molecular structure of ((IPr)Ni)2(μ-CO)(μ-η2,η2-CO2) [68].

In 1998, Gibson et al. reported [69] the synthesis of ruthenium μ2-η2-type CO2bridged complexes from the reaction of cis-Ru(phen)2(CO)(CHO)(PF6) with O2 in dimethyl ether (DME)/H2O: (Ru(phen)2(CO)(CO2)Ru(phen)2(CO))(PF6)2, and of the related complex (Ru(phen)2(CO)(CO2)Ru(bpy)2(CO))(PF6)2. The latter complex was structurally characterized as having the following values: the OCO angle was 122.3 °; the CO bonds were 1.235 and 1.319 Å; and the IR bands for the carboxyl bridge were located at 1499 and 1183 cm−1. Another example of a multisite bridging CO2 ligand, starting from a carbonyl complex, was reported in 2001 by Chi et al., with the preparation and characterization of two CO2-bridged osmium cluster complexes [70]. These authors reported the synthesis of Os4(μ-H)(μ-CO2)(thd)(CO)13 (tdh = 2,2,6,6-tetramethyl-3,5-heptandionate) by the direct treatment of Os3(CO)12 with a β-diketone molecule (thd)H. The XRD study showed how a unique CO2 ligand could bridge a triosmium metal fragment, Os3(μ-H)(CO)10 through the C and O atoms, and a monometallic osmium moiety, Os(CO)3(thd) through one O atom. This was a different pattern from that of cluster complexes with a bridged carboxylate ligand, where the CO2 was linked to the nearby Os–Os edge via two direct Os–O σ-bonds. The OCO angle was 118 °, which was close to the value encountered in CO2− moieties, while the average CO bond length was 1.27 Å. In 2004, Wang et al. showed that air-stable aminorhenium–CO2 complexes could be easily obtained in good yields from the corresponding carbonyl complexes η5:η1-C5H4CH2CH2NR(CH3)Re(CO)2, by oxidation with peroxy-acids [71]. The binding mode of the CO2 ligand to the metal was the “side-on” type, with a OCO angle of 133 °, and CO bond lengths of 1.19 and 1.23 Å, similar to other η2-CO2 complexes reported elsewhere (Figure 4.7). On this basis, a reaction mechanism

4.4 Reactivity of CO2 Complexes

Figure 4.7

ORTEP view of η5:η1-C5H4CH2CH2NR(CH3)Re(CO)(η2-CO2) complex [71].

was proposed whereby a transient Re(CO)2(O−) intermediate underwent intramolecular cyclization to produce the CO2 complex. The ﬁnal example is related to the redox transformations of CO2 in biological cycles. Carbon monoxide dehydrogenases (CODHases) are the biological catalysts for the reversible oxidation of CO to CO2, with water as the source of oxygen: CO + H2O → CO2 + 2e− + 2H+. In 2007, Dobbek et al. [72] reported crystal structures of anaerobic CODHases containing Ni-, Fe-, S- cluster in three different states. In the intermediate structure, CO2 was seen to act as a bridging ligand between Ni, completing the square-planar coordination of Ni2+, and the asymmetrically coordinated Fe, replacing a water hydroxo-ligand bound to Fe2+ in the two other states. In the cluster–CO2 complex, Ni acted as the Lewis base, and Fe was the Lewis acid that stabilized the negative charge on oxygen. CO2 acted as a η1-OCO end-on ligand at Fe, and as a η1-C ligand at Ni, which resulted in a μ2-η2 binding mode of CO2 bridging the Ni–Fe site. The oxygen atoms were also H-bonded to two protein fragments (Lys and His), while the OCO angle was equal to 133 °.

4.4 Reactivity of CO2 Complexes

It is clear that CO2 is in an “activated state” when the bonding involves the central carbon. The activation is evident from structural data such as the bent OCO moiety

69

70

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

and the increased CO distances, and from spectroscopic data such as lower frequency vibration modes in IR spectra and low ﬁeld shifts in 13C NMR spectroscopy. It can be expected that this “activation” would also be reﬂected in the reactivity of coordinated CO2. However, the relationships between these model complexes and their reactivity are not yet understood, nor the structural and electronic properties of metal–CO2 moieties that are required for catalytic activity. Very few examples of reactions of “true” CO2 complexes are known. Among them, the cleavage of a CO bond is often observed and yields metal oxides that can react further with CO2, leading to a net conversion of two CO2 molecules to CO and CO2− 3 . The same conversion can occur via the intermediate coupling of two CO2 molecules at a metal center, as described previously for Carmona’s Mo complex [12, 67]. Coordinated CO2 exhibits an enhanced reactivity at oxygen towards electrophiles, leading to metallacarboxylic acids and their derivatives. Electrophilic attack at oxygen is also the prevalent reactivity of η1-CO2 complexes and CO2-bridged complexes with alkali cations. Many stoichiometric and most catalytic reactions involving CO2 activation proceed via the formal insertion of CO2 into highly reactive M–L bonds, with the formation of new C–L bonds (where L = ligand). These reactions might not necessarily require strong coordination of CO2, as in the stable complexes described in Section 4.3.2, but are generally initiated by a nucleophilic attack of L at the Lewis acidic carbon atom of CO2. A weak interaction between the metal and the lone pairs of one oxygen atom or the π-C=O bond of CO2 may also play a role in supporting the insertion process. Three types of possible reaction of coordinated CO2, illustrated with recent examples of the past decade, are described in the following sections. These include CO bond cleavage and oxygen transfer, reactions with electrophiles, and reactions with nucleophiles. 4.4.1 C–O Bond Cleavage and O Transfer

A reaction that may be characteristic of η2- and η1-complexes formed from transition metals is thermolysis with cleavage of one CO bond resulting in the loss of CO and formation of an oxo compound. The conversion of CO2 into the coordinated CO and phosphine oxide was observed by Ohnishi et al. [73] in the reaction of the Mo(0) complex Mo(P4)(dppe) with CO2 to give a M(0) carbonyl complex fac-Mo(CO)(η3-P4=O)(dppe), where the O abstraction from CO2 by one terminal P atom in P4 (meso-o-C6H4(PPhCH2CH2 PPh2)2) takes place to give the dangling P(=O)Ph2 moiety together with the coordinated CO. The proposed mechanism involves the initial displacement of one Mo–P bond to give a η2–CO2 complex Mo(η2-CO2)(η3-P4)(dppe) that undergoes O atom abstraction by the dangling P atom of the η3-P4 ligand. Later, Srivastava et al. [74] reported the catalytic activities of several Cu and Mn aza complexes for cyclic carbonate synthesis from CO2 and epoxides. These authors showed that the reaction of CO2 with the catalyst system generated two new IR bands

4.4 Reactivity of CO2 Complexes O–

O C +

R

O

Cu

CO2 Cu

O R O O

DMAP

Cu

DMAP

O R O Cu

O

– O O

Cu

+ C

O R

R

Cu

DMAP

O

Cu R

DMAP

O Cu

O

C

Cu

DMAP

O Figure 4.8 Possible reaction mechanism for cyclic carbonate synthesis from CO2 and epoxides on Cu aza complexes [74].

at 1716 and 1225 cm−1, due to the formation of an activated Cu complex in a pentacoordinated structure with CO2 bound to the metal in a η1-C mode. The activated CO2 complex then attacked the activated epoxide at the less-hindered carbon, leading to the formation of a dimer intermediate species which yielded the cyclic carbonate (Figure 4.8). The nature of the ligands, and of their substitutions, inﬂuenced the lability of the metal–CO2 bonding, and the differences in the catalytic activities were correlated with the mode and strength of CO2 binding. The metal–CO2/epoxide bonding should neither be too strong nor too weak since, by a suitable ﬁne-tuning of the ligand system, it would then possible to develop a highly active catalyst. One of the most interesting recent examples of oxygen transfer reaction has been the studies conducted by Meier and colleagues [75–77] of the activation of CO2 by uranium complexes. Certain ligands form multiple bonds to uranium, and small inert molecules such as CO2 become reactive when coordinated to the metal. The ability of U to use its outermost f electrons for binding ligands enables it to catalyze reactions that should be impossible with “conventional” transition metal catalysts. The oxygen transfer to a metal center, with formation of the O=M(CO) moiety, may be driven by the ancillary ligands, as shown recently in the case of the U(III) complex (see Refs [63] and [75]). A new rather stable complex

71

72

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

LU + CO2 1

LU=

O

N O UIII N N O

LU–(CO2)–UL 3

CO + LU–O–UL 2 LU + CO 1

LU–CO–UL 4

Figure 4.9 Molecular representation of ((AdArO)3tacn)U(CO2) with core structure and geometric parameters in Å and degrees, and possible mechanism of CO2 reduction by a less-bulky UIII complex [75].

((AdArO)3tacnUIII)=O=C•–O−, where CO2 is coordinated to U in a η1-O mode and reduced by one electron to a radical anion, has been described in Section 4.3.2.1. However, depending on the steric hindrance of the ligand around the U center, a further reaction can occur. By replacing the adamentane ligand (Ad) with a lesslarge tert-butyl ligand (t-Bu), an initial formation of a CO2 adduct will occur, following which a second equivalent of (t-BuArO)3tacnUIII enters to complete the reduction process, to form a bridging oxide complex ((t-BuArO)3(tacn)U-O-U(tacn) (t-BuArO)3) and releasing CO (Figure 4.9). 4.4.2 Reactions with Electrophiles

Although the reports of chemical reactivity of coordinated CO2 ligand in wellcharacterized complexes are still very limited, several examples have been described showing that the oxygen of the η2- or η1-bonded coordinated CO2 can undergo an electrophilic attack by protons or other similar reagents. For example, the CO2 complex of iron(0), Fe(CO2)(depe)2, which has a trigonal bipyramidal geometry with a side-on bonded CO2, reacts with electrophiles such

4.4 Reactivity of CO2 Complexes

as MeI and MeOTf to give Me2O and the corresponding cationic iron(II) carbonyl complexes FeX(CO)(depe)2)+X−, with X = I or OTf [78]. The reaction of Fe(CO2) (depe)2 with R3SnCl in diethyl ether at 195 K gives the iron carboxylate complexes FeCl(CO2SnR3)(depe)2 with R = Me or Ph, where the CO2 moiety bridges the Fe and Sn atoms in a μ-η1(C):η2(O,O)–CO2 fashion. The formation of the iron carboxylate and the cationic carbonyl complexes can be interpreted by the following mechanism. First, the electrophile interacts directly with one of the oxygen atoms of the CO2 ligand to give a metal carboxylate FeC(O)OR. Then, with organotin chlorides, a stable heterodinuclear carboxylate FeCO2Sn is formed. With stronger electrophiles such as RX, decarboxylation takes place to give an alkoxocarbonyl iron Fe+(CO)(OR)X− that further reacts with another electrophile RX to form the corresponding ether R2O and a halogenocarbonyliron(II) halide XFe+(CO)X−. Another interesting example is the reaction of hydrosilylation of CO2. A recent report by Deglmann et al. [79] has described, both experimentally (by using in situ IR spectroscopy) and theoretically (by running DFT calculations with a Me3SiH model), the catalytic hydrosilylation of CO2 by Me2PhSiH with ruthenium nitrile complexes mer-(RuX3(MeCN)3) and cis/trans-(RuX2(MeCN)4), with X = Br, Cl. The key steps of the reaction were shown to be the transfer of the Me3Si moiety to a coordinated halide ligand, resulting in a LnRuH–(XSiMe3) intermediate, CO2 coordination to Ru in a side-on fashion, Me3Si transfer to CO2, and reductive elimination of the formoxysilane product Me3SiOCHO (Figure 4.10). The general orders of activity for CO2 hydrosilylation were found to be RuClmLn > RuBrmLn and RuIII > RuII. The highest point on the energy pathway was found to be the silyl transfer from the R3SiCl ligand to the oxygen atom of the η2-bound CO2. In 2001, Matsubara et al. [80] performed a theoretical study of the mechanism of the hydrido migration to CO2 of Fe, Ru, and Os complexes containing a protonated amine arm by DFT calculations, asking the question: which pathway did the reaction take? This might be an abstraction of the hydrido ligand by an incoming CO2 without direct coordination of CO2 to the metal atom (pathway a, where the amine arm promotes the electrophilic attack of CO2 and stabilizes the product), or the insertion of CO2 into the M–H bond with the previous η2-CO coordination of CO2 to the metal atom (pathway b). Matsubara and coworkers concluded that the reactivity was reduced in the order Fe > Ru > Os and that, in the case of Fe, pathway b was preferred, whereas pathway a was more favorable for Ru and Os, by a careful analysis of geometries, charges and energy barriers along the reaction paths. This showed how charge distributions and bond populations were important clues for understanding the mechanisms of reactions initiated by electrophilic or nucleophilic attacks. 4.4.3 Reactions with Nucleophiles

Very few reactions of CO2 complexes with nucleophiles have been reported. One of the rare examples of nucleophilic attack at CO2 has been provided by Aresta et al. [81], in a study of the coordination chemistry of phenoxide ligands to MnII

73

74

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2 L L L RuI CI CI A L -L

L L L RuIII CI H CI O C O +L SiMe3 I

L L RuI CI L CI -Product

H SiMe3 L L RuII CI L CI C

B

H L C O L RuII CI CI O H SiMe3

‡ SiMe3 L H L RuII CI L CI TS1

L H CI L Ru O CI C O SiMe3

‡ L H SiMe3 L RuII CI L CI D

TS3

-L H SiMe3 L I CI Ru L CI E

L H L RuIV CI O CI C ‡

O G

SiMe3

L H L Ru CI O CI C SiMe3 O TS2

+CO2 L H SiMe3 L RuII CI C O CI O F

Figure 4.10 Calculated mechanism of the hydrosilylation of CO2 involving a neutral RuII nitrile complex as catalyst [79].

because of the possible implication of MnII–phenoxide complexes as intermediates in the phenylphosphate carboxylase enzyme, a protein which catalyzes the selective carboxylation of phenylphosphate to 4-OH-benzoic acid using CO2. Aresta and colleagues reported the synthesis and the characterization of mixed-sandwich complexes with Cp and phenate as π-ligands CpMn(η5-ArO), with ArO = 2,6substituted phenoxide being π coordinated to the CpMn+ moiety. When these complexes were exposed to CO2, the formation of 4-OH-3,5-substituted benzoic

4.5 CO2 Complexes as Reaction Intermediates in CO2 Utilization Processes

acid was observed, because the π coordination forced electron dislocation in the 4-position to facilitate the nucleophilic attack of the ring carbon of the phenoxide on CO2. This produced the ring carboxylation, whereas the μ-O-bonded phenols (CpMn(μ-OAr)(THF))2 underwent CO2 insertion into the Mn–O bond. Another example involving Aresta’s complex (PCy3)2Ni(η2-CO2) is the Wittig reaction, observed for this compound by Gong et al. [82]. The so-called Wittig reaction is a (2+2) cycloaddition mechanism between strong nucleophiles such as trialkyl phosphorous ylides R3P=CH2 and organic ketones or aldehydes to form alkenes. Actually, when CO2 is bubbled in THF with Me3P=CH2, free ketene CH2=C=O is formed, but this immediately dimerizes to form methylene-βpropiolactone. Ketenes can be stabilized by coordination on transition metal complexes in a η2-C,O (for early transition metals) or η2-C,C (for late transition metals) manner. In this study, Aresta’s complex was reacted with Me3P=CH2 in toluene at 253 K for 2 h, and a unique nickel ketene complex (Cy3P)2Ni(η2-C,O)–CH2=CO was isolated and characterized (1611, 1570 cm−1). The reaction mechanism could be a nucleophilic attack of coordinated CO2 by the ylide; alternatively, it is also possible that the CO2 had dissociated from Ni, reacted with the ylide in the solvent, and that the ketene further coordinated to the nickel. Nevertheless, the unusual η2-C,O ketene bonding mode observed on Ni was more in favor of a nucleophilic attack of coordinated CO2. In contrast, by using kinetic measurements, Konno et al. [83] showed that the nucleophilic attack of the hydride ligand of Ru(tpy)(4,4′-X2bpy)H+ (X = H, MeO) to the carbon atom of CO2 was the rate-determining step for the formation of the formate complex Ru(tpy)(4,4′-X2bpy)(OCHO)+, while the η1-O or η2-C,O coordination of CO2 on Ru should be less important, if at all, in the transition state. Another recent example of the question of the formation of intermediate metal– CO2 complexes in these reactions was the theoretical study by Ohnishi et al. [84] of the hydrogenation of CO2 to formic acid by RuII catalysts. In the presence of water, there was no direct metal coordination of CO2, but formation of adducts in which the C and O atoms of CO2 interacted with the H (hydride) ligand and the H atom of H2O: cis-Ru(H)2(PMe3)3(H2O)(CO2). In the absence of water molecules, CO2 directly coordinated to the Ru center to afford Ru(H)2(η2-CO2)(PMe3)3. The Ru-(η1-formate) intermediate was then produced via CO2 insertion, instead of a nucleophilic attack of the H ligand to CO2 in the presence of water. As a result, the activation barrier was higher, the exothermicity lower, and a back reaction could occur. Hence, this is an example where the formation of a metal–CO2 complex is not favorable to the further reaction.

4.5 CO2 Complexes as Reaction Intermediates in CO2 Utilization Processes

From the point of view of synthesis, the most interesting reactions of CO2 result in the formation of new bonds between the carbon atom and a second group, L. Many reactions of this type are known to proceed in the presence of stoichiometric

75

76

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

or catalytic amounts of transition metal complexes. In principle, these conversions can occur via one of the three following pathways:

• • •

Attack of the reactant at the coordinated CO2: LnM(CO2) + L → LnM-O-C(O)L (see Section 4.3) Reaction of free CO2 with a metal complex: LnM(L) + CO2 → LnM-O-C(O)L Simultaneous coordination of CO2 and L at the metal: LnM(CO2)(L) → LnM-OC(O)L

The thermal lability of CO2 complexes inﬂuences their further chemical reactions. The η1 and η2 complexes are very prone to dissociate the coordinated CO2. For example [85], when Aresta et al. studied the mechanism of formation of the peroxocarbonate complex (PCy3)2Ni(CO4) from solid (PCy3)2Ni(CO2) and dioxygen, it was proved – via 13C and 31P NMR spectra at variable temperatures, together with a FTIR study using 18O2-, 13CO2-, and C18O2-labeled compounds – that the reaction involved CO2 de-coordination, O2 coordination, and reinsertion of CO2 into the O–O bond of the reactive newly formed oxygen complex (PCy3)2Ni(O2). When CO2 is bound between two metal centers, the ones that lose CO2 most readily are the μ2-η2 and μ2-η3 complexes, in which one or two of the carboxylate oxygen atoms is (are) bound to a main group atom. These complexes are often intermediates in photochemical and electrochemical reduction reactions of CO2 to CO. Oxidative coupling reactions, reduction reactions and catalytic processes are described in other chapters of this book and in recent reviews [86–88]. In this section, attention is focused only on some recent examples of synthetically useful processes where new C–L bonds are formed, and where the formation of a metal– CO2 complex as an intermediate has been suspected or demonstrated by spectroscopic evidence and/or calculations. The peculiar case of a few bioinspired reactions is also cited. 4.5.1 Oxidative Coupling Reactions

Most oleﬁns undergo oxidative coupling reactions with CO2 in the presence of transition metal centers such as Ni, Fe, Mo, or W, to produce lactones, carboxylic acids, and acrylate derivatives. The reaction of alkenes or alkynes with CO2, mediated by nickel complexes, was reported by Hoberg et al. during the 1980s (see Refs [89, 90]), and supplemented by theoretical studies (see Refs [91, 92]). Since then, studies of oxidative coupling reactions of CO2 with unsaturated compounds at a transition metal center have received much attention, though attempts to overcome the noncatalytic nature of the coupling reaction have proven unsuccessful. In this section, attention will be focused only on recent theoretical studies, the aim being to understand what these reaction mechanisms are, and why so far it has been difﬁcult to obtain catalytic cycles. Detailed theoretical studies have been conducted by Papai et al. on acrylate formation mechanism via Mo- [93] and Ni-assisted [94] C–C coupling between CO2

4.5 CO2 Complexes as Reaction Intermediates in CO2 Utilization Processes

77

and C2H4 from DFT calculations. In the reaction of trans-Mo(C2H4)2(PMe3)4 with CO2, one of the key intermediates was shown to be a complex involving both coordinated CO2 and C2H4 ligands. The coupling took place in a coplanar arrangement via a ﬁve-center transition state, and resulted in a near-planar metalla-lactone species. Thus, the coordination of CO2 to the metal center was seen to be a necessary step to achieve the coupling, and was made possible by one phosphine dissociation. The electronic effects of the phosphine ligands were also seen to be important, as the C–C coupling was facilitated by σ-donation, because the metal atom acted as a basic center in the coupling process via metal → CO2 and metal → C2H4 electron donation mechanisms. The overall reaction was found to be exothermic (by 27 kcal mol−1), with all intermediates lying below the reactants level and separated by reasonably low energy barriers; the highest barrier (11 kcal mol−1) corresponded to the C–C coupling step. In contrast, in the reaction Ni (cdt ) + L2 + CO2 + C2H4 → L2Ni (CH2CH2C (O) O) + cdt cdt = cyclododecatriene L2 = bipyridine ( bpy ) or bis (dicyclohexylphosphino ) ethane ( dcpe) the same authors [94] showed that the C–C bond formation occurred in a single step by the reaction of a nickel–ethylene complex with an incoming CO2 molecule, and that the lowest energy pathway did not involve CO2 coordination, as suggested previously [91] (Figure 4.11). Although the L2Ni(CO2)(C2H4) complex has been identiﬁed as a minimum on the potential energy surface of the reaction, the barrier to dissociation is extremely low, and so its presence in the reaction mixture would be rather unlikely. The nature of the ancillary ligands was also shown to be important, as it affected the relative stability of the minima associated with the C–C bond formation, and, therefore, the related activation barriers. As noted previously, this was related to the σ-donating ability of the chelating ligand, and, for example, for L2 = dcpe, the C–C bond formation was reversible. This reaction was recently reinvestigated [95] in order to propose a calculated route for the catalytic production of acrylic acid. Similar to previous studies [94],

N

N +

Ni

N

N

N

+ C2H4

Ni

N

N

Ni + CO2 N

N

N

N C O

O

Ni C O O Figure 4.11

N Ni

Ni

Mechanistic details of Ni(0)-assisted oxidative coupling of CO2 with C2H4 [94].

O O

78

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

Figure 4.12 [95].

Proposed catalytic cycle for the coupling of CO2 and C2H4 to produce acrylic acid

the authors concluded that the formation of the intermediate nickelacycle occurred in a single step via the reaction of nickel-coordinated ethylene with an incoming CO2 molecule (Figure 4.12). The whole calculated catalytic cycle was shown to contain three main barriers: nickelacycle formation (122 kJ mol−1); β-H elimination (147 kJ mol−1); and the reductive elimination of acrylic acid (104 kJ mol−1), that were not expected to hinder the progress of the reaction. Likewise, the catalyst deactivation reaction should not be the cause of noncatalytic activity. It was suggested that the lack of catalytic activity was more likely due to the free energy of the reaction CO2 + C2H4 → CH2=CHCO2H, calculated to be unfavorable by 43 kJ mol−1. Although, the overall thermodynamics created an unfavorable equilibrium, it was possible that the catalytic formation of acrylic acid would occur at the start of the reaction that led to undetectable amounts. Possible solutions may include preventing the reaction from reaching equilibrium through removal of the product. Recently, Aresta et al. [96] also reported a new synthetic approach to acrylates, starting from a preformed Pd-COOMe moiety, and the ﬁrst formation of ethyl acrylate from ethene and CO2 on a palladium complex (L-L)PdH+. Although, at

4.5 CO2 Complexes as Reaction Intermediates in CO2 Utilization Processes

present the mechanism involved remains unknown, theoretical calculations are currently in progress in an effort to elucidate its details. Factors controlling the regio- and stereo-selectivity of metal-centered CO2 reactions are also often not understood. Theoretical studies performed using DFT calculations on coupling reactions of CO2 with terminal and internal alkynes mediated by Ni(0) complexes were also recently reported [97]. Here, the authors showed that the coupling reaction mainly proceeded through an associative mechanism in which a direct electrophilic attack of CO2 at the π⊥ bond of the η2-coordinated alkyne occurred. Coupling reactions with terminal alkynes provide the same regioselectivity, whatever the alkyne electronic properties are, while electronic factors affect only the reaction barriers. For substituted internal alkynes, both steric and electronic factors favor coupling between CO2 and the substituted carbon. Again, the nature of the ancillary ligand is important, and a more σ-donating ligand, such as dbu (diazabicycloundecene), has been found to be a better ligand than bpy – that is, it leads to lower energy coupling barriers. 4.5.2 Reduction Reactions

By employing enzymes, Nature easily uses CO2 as a one-carbon building block for the synthesis of organic molecules. Yet, the reduction of CO2 to CO requires a large amount of energy, and even in the presence of a strong reducing agent, overcoming the O=CO bond enthalpy (532 kJ mol−1) often presents kinetic difﬁculties. Some metal complexes (such as the uranium complexes described above [75–77]) abstract oxygen from CO2, but the metal–oxygen bonds formed are strong, and catalytic cycles are difﬁcult to achieve. Recently, Sadighi et al. [98] reported a copper–boryl complex (NHC)Cu(Bpin) (NHC = N-heterocyclic carbene; pin = pinacolato) that abstracts oxygen from CO2 and readily undergoes subsequent turnover: L-Cu-Bpin + CO2 → L-Cu−O-Bpin + CO L-Cu−O-Bpin + ( pinB)2 → ( pinB)2 O + L-Cu-Bpin Shortly afterwards, Zhao et al. proposed a mechanism based on DFT calculations for this reaction [99]. They showed the ﬁrst step of the reaction to be a η2-CO2 coordination to the copper center, that activates signiﬁcantly the C=O bond, with an electron transfer of 0.53e from the metal fragment to the CO2 moiety, suggesting a signiﬁcant back-bonding in the complex. The Cu–B σ-bond was also involved in a bonding interaction with an empty π* orbital on the η2-CO2 ligand. The coordinated CO2 then inserted into the Cu–B bond with a small barrier of 3.8 kcal mol−1, giving a Cu–O–C–B linkage and boryl migration from C to O, followed by a σ-bond metathesis between pinB–Bpin and (NHC)Cu(OBpin) (Figure 4.13). The boryl migration from C to O, which releases CO, was shown to be the rate-determining step of the reaction. The electron richness of the Cu–B bond, due to the low electronegativity of B, gave rise to a small CO2 insertion barrier because the

79

80

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

Figure 4.13

Mechanism of reduction of CO2 to CO catalyzed by Cu(I) boryl complex [99].

back-bonding interaction between the Cu–B bond and the CO2 carbon was recognized as being important in the insertion process. A catalytic reduction of CO2 to CH4 was described in 2006 by Matsuo et al. [100], using hydrosilanes: CO2 + 4 Si–H → CH4 + 4 Si–O, catalyzed by zirconium–borane complexes LZr(CH2Ph)(B(C6F5)4). These authors showed that the combination of both zirconium and boron provided the catalytic activity. In the absence of Zr, the reduction of CO2 was not observed, which suggested that the coordination of CO2 to the Zr center took place prior to the actual reduction. This coordinated CO2 would then be more reactive towards hydrosilylation so as to yield the initial product CH2(OSiR3)2, although the nature of the counterion was also shown to be critical during the course of the reaction. The reductive disproportionation of CO2 to give CO and CO2− 3 , promoted by electron-rich transition metals [9, 101] or to produce C2O2− using lanthanides such 4 as Sm [102, 103], has been observed in a number of systems. For example, Allen et al. have recently isolated an iron carbonyl complex with a bicarbonate counterion (Fe(dmpe)2(CO)H)(HCO3) with dmpe = 1,2 bis(dimethylphosphino)ethane, resulting from the iron(II)-mediated reductive disproportionation of CO2 [104], although the reaction mechanism is still speculative. In a related system, Karsh et al. [9] had observed, during the reaction of CO2 with Fe(PMe3)4, the formation of a side-on complex Fe(CO2)(PMe3)4 and of a carbonyl carbonate species Fe(PMe3)3(CO)(CO3). With regards to actinides, Summerscales et al. [105] have also reported the reductive disproportionation of CO2 to carbonate and squarate products, using COTR/ CpR′ mixed-sandwich U(III) complexes (COT = cyclooctatriene, Cp = cyclopentadienyl, R = SiiPr3, R′ = H, Me). U(III) systems have been shown previously to reduce CO2 to a linear η1-O-coordinated mono-anion [63], and to doubly reduce to give O2− salts plus CO [75, 76]. In this case, the reactions can be considered

4.5 CO2 Complexes as Reaction Intermediates in CO2 Utilization Processes R

R

10

U

+

O

8 CO2

2 R = SiiPr3

d8-toluene -30

25°C

R

R

R 4

U

R

O O

C O

U

1

O

O

U O

R

5b

+

R

O

4

U

R R

Figure 4.14 Reductive disproportionation of CO2 to carbonate and squarate products using a mixed-sandwich U(III) complex [105].

as successive 2e- reductions, although the reaction mechanism has not been described: 8 CO2 + 8 e− → 4 CO23 − + 4 CO + 2 e− → 4 CO23 − + C4O24− The bridging carbonate has an unusual μ-η1:η2-mode of coordination to U atoms with a ﬂuxional structure, whereas the squarate species has a μ-η2:η2 structure (Figure 4.14). This is the ﬁrst example of the formation of a carbocycle derived purely from a CO2 carbon source. 4.5.3 Catalytic Processes

Recent studies of three different catalytic reactions involving CO2 complexes as reaction intermediates are listed below. In the ﬁrst reaction, Choi et al. [106] studied the oxidative addition of benzene to RhCl(PMe3)3 for the photochemical formation of mer-Rh(C6H5)(H)Cl(PMe3)3. Interestingly, the presence of an excess of CO2 in the solution promoted the

81

82

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

Figure 4.15 Possible mechanisms for disubstituted urea synthesis (I) and cyclization reactions of epoxides (II) [107].

reaction. The authors suggested possible explanations that included: (i) the formation of a small amount of CO and O=PMe3, followed by the preferential photoreaction of RhCl(CO)(PMe3)2; or (ii) the reversible coordination of CO2 to the Rh center, resulting in the stabilization of productive intermediates such as RhCl(C6H5)(H) (CO2)(PMe3)2. No insertion of CO2 into the Rh–H or Rh–C bond was observed, but the prolonged photoreaction produced a small amount of a phenyl(carbonato) rhodium complex, due to the disproportionation of CO2 to CO2− 3 and CO. The second example dealt with the reactions of amines or epoxides with CO2 on supported nanogold catalysts [107]. These authors showed that polymer-immobilized nanogold catalysts had a very high catalytic activity for the activation of carbon dioxide, with a turnover frequency (TOF) in excess of 50 000 mol mol−1 h−1 for the synthesis of cyclic carbonate, and a turnover frequency of product (TOFP) of 3000 mol mol−1 h−1 for the synthesis of disubstituted ureas. Whereas, the reaction mechanism is not yet clear, it was suggested that the coordination onto the gold particles might represent the key step, and that the activation of CO2 over polymersupported nanogold clusters could be universal (Figure 4.15). The third reaction was related to the hydrocarboxylation of allenes with CO2, catalyzed by a tridendate silyl pincer-type palladium complex [108]. In this reaction, a σ-allyl palladium species (via hydropallation of allene) was formed, and its trigonal bipyramidal geometry allowed the facile coordination of CO2 (presumably in a side-on fashion) and a following nucleophilic addition to realize the carboxylation of allene (Figure 4.16). This reaction proved to be very attractive not only as a CO2 ﬁxation reaction, but also as a general method for the synthesis of βγ-unsaturated carboxylic acids. 4.5.4 Bioinspired Reactions

Among the enzymes that occur in biological systems and which utilize CO2 as a source of carbon, some contain metal atoms as the active center, where CO2 is converted (as in RuBisCO or biotin-dependent carboxylases). However, some of these enzymes do not require metal, or, if so, it does not interact directly with CO2.

4.5 CO2 Complexes as Reaction Intermediates in CO2 Utilization Processes

Figure 4.16 Utilization of silyl pincer-type Pd complex for the hydrocarboxylation of allenes with CO2 [108].

O O

H L

O Zn

L

H + CO2 L

1

L

O Zn

L

H O C O L

L

O

Zn

L L 3 (TS)

2

O H

O Zn

O L

bond rotation vs. proton transfer

L

L 4

H

O

O L

O Zn L

L

5

+ H2O - H+/HCO3Figure 4.17

The mechanism of CO2 ﬁxation by carbonic anhydrase [110].

In natural processes, metal ions are often in high oxidation states (2 or 3), whereas in chemical systems the metals are in low oxidation states (0 or 1). This fact inverts the role of the metal center, such that it acts as a one-electron sink in a natural system, but as a nucleophile in an artiﬁcial ones (see other chapters of this book and the review by Aresta et al. [109]). Nevertheless, important biochemical processes such as the reversible enzymatic hydration of CO2, or the formation of metal carbamates, may serve as natural models for many synthetic purposes. Starting from the properties of carbonic anhydrase (a zinc metalloenzyme that performs the activation of CO2), Schenk et al. proposed a review [110] of perspectives to build biomimetic chemical catalysts by means of high-level DFT or ab initio calculations for both the gas phase and in the condensed state. The ﬁxation of CO2 by Zn(II) complexes to undergo the hydration of CO2 (Figure 4.17); the use of Cr, Co, or Zn complexes as catalysts for the coordination–insertion reaction of CO2 with epoxides; and the theoretical aspects of carbamate synthesis, especially for the formation of Mg2+ and Li+ carbamates, are discussed in the review of Schenk

83

84

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2

et al. The aim was to suggest novel synthetic pathways that might result in modiﬁed metal complexes with catalytic activities comparable to those of natural precursors. The mechanism of the formation of peroxocarbonates LnM(CO4)Xm, and their reactivity as oxygen-transfer agents, is important owing to their potential applications as oxidants of phosphines, oleﬁns, methylene groups, and also as compounds mimicking monooxygenases. Aresta et al. have shown [111], via extended FTIR experiments with labeled 13CO2 and C18O2, that the formation of rhodium peroxocarbonate from CO2 and RhCl(η2-O2)(P)3, with P = PEt2Ph or PEtPh2, proceeds through O–O bond cleavage and CO2 insertion into the O–O bond, rather than the insertion of CO2 into a Rh–O bond, allowing the synthesis of asymmetric 16 O–18O moieties. The further O-transfer to ancillary phosphine ligand to give R3P=O selectively involves the Rh-linked O atom of the peroxo group of RhCl(CO4) (P)3. In this case, no evidence has been found for a direct coordination of CO2 on to the metal center during the reaction.

4.6 Conclusions

Knowledge of CO2 binding and activation has increased dramatically since the ﬁrst stable CO2 complex Ni(CO2)(PCy3)2 was structurally characterized more than three decades ago [7, 8]. Different binding modes of CO2 in mononuclear and polynuclear metal complexes have been characterized. In 2004, the ﬁrst stable end-on coordinated CO2 ligand was characterized in a uranium complex [63], followed in 2005 by a novel Al–Mg complex containing CO2 in a linear μ(O,O′) bonding mode [64]. The major improvement that have been made in theoretical calculations and in spectroscopic methods now allow the identiﬁcation and characterization of stable or unstable complexes, together with the elucidation of reaction mechanisms. Less attention has been paid, however, to CO2 organometallic chemistry during the past decade. Whilst many reduction or coupling reactions are known to proceed in the presence of stoichiometric or catalytic amounts of transition metal complexes, very few examples remain where the formation of a metal–CO2 complex has led to an effective, catalytic reduction reaction of CO2. Carbon dioxide complex photoactivation also represents an attractive route to CO bond cleavage, coupled with O-atom transfer. However, progress in the area of CO2 utilization requires a better understanding of the reaction mechanisms, of the thermodynamics of reaction intermediates, and of structure–reactivity relationships. Although today, more is understood about CO2 coordination, knowledge of the coordination site requirements in various CO2 reactions remains poor, as does that of cooperative interactions with co-ligands. Consequently, systematic studies of this important mechanistic aspect are required, using physico-chemical techniques and computer modeling. It is important to elucidate reaction mechanisms at the molecular level, and topics such as acrylic synthesis from ethylene and CO2

References

or the carboxylation of CH bonds are of major interest. The effective activation of CO2 by metal complexes is a goal that may be difﬁcult to reach, but which represents an exciting research area in organometallic chemistry.

Acknowledgments

J.M. thanks Dr Imre Papai, of the Chemical Research Centre, Budapest, for helpful discussions, and for providing Figures 4.1 to 4.3.

References 1 Pandey, K.K. (1995) Coord. Chem. Rev., 140, 37–114. 2 Gibson, D.H. (1996) Chem. Rev., 96, 2063–2095. 3 Leitner, W. (1996) Coord. Chem. Rev., 153, 257–284. 4 Gibson, D.H. (1999) Coord. Chem. Rev., 185–186, 335–355. 5 Yin, X.L. and Moss, J.R. (1999) Coord. Chem. Rev., 181, 27–59. 6 Behr, A. (1988) Carbon Dioxide Activation by Metal Complexes, WileyVCH Verlag GmbH, Weinheim, Germany. 7 Aresta, M., Nobile, C.F., Albano, V.G., Forni, E., and Manassero, M. (1975) J. Chem. Soc. Chem. Commun., 636–637. 8 Döhring, A., Jolly, P.W., Krüger, C., and Romano, M.J. (1985) Z. Naturforsch., 40b, 484. 9 Karsch, H.H. (1977) Chem. Ber., 110, 2213–2221. 10 Calabrese, J.C., Herskovitz, T., and Kinney, J.B. (1983) J. Am. Chem. Soc., 105, 5914–5915. 11 Bristow, G.S., Hitchcock, P.B., and Lappert, D.M. (1981) J. Chem. Soc. Chem. Commun., 1145–1146. 12 Alvarez, R., Carmona, E., GutierrezPuebla, E., Marin, J.M., Monge, A., and Poveda, M.L. (1984) J. Chem. Soc. Chem. Commun., 1326–1327. 13 Gambarotta, S., Floriani, C., ChiesiVilla, A., and Guastini, C. (1985) J. Am. Chem. Soc., 107, 2985–2986. 14 Fachinetti, G., Floriani, C., and Zanazzi, P.F. (1978) J. Am. Chem. Soc., 100, 7405–7407.

15 Tanaka, K. and Ooyama, D. (2002) Coord. Chem. Rev., 226, 211–218. 16 Jegat, C. and Mascetti, J. (1991) New J. Chem., 15, 17–20. 17 Freund, H.J. and Roberts, M.W. (1996) Surf. Sci. Rep., 25, 225–273. 18 Almond, M.J. and Downs, A.J. (1989) Spectroscopy of Matrix Isolated Species, John Wiley & Sons, Ltd, Chichester, England. 19 Jacox, M.E. and Milligan, D.E. (1974) Chem. Phys. Lett., 28, 163–168. 20 Kafaﬁ, Z.H., Hauge, R.H., Billups, W.E., and Margrave, J.L. (1984) Inorg. Chem., 23, 177–183. 21 Hauge, R.H., Margrave, J.L., Kauffmann, J.W., Rao, N.A., Konarski, M.M., Bell, J.P., and Billups, W.E. (1981) J. Chem. Soc. Chem. Commun., 1258–1260. 22 Manceron, L., Loutellier, A., and Perchard, J.P. (1985) J. Mol. Struct., 129, 115–124. 23 Hwang, D.Y. and Mebel, A.M. (2000) Chem. Phys. Lett., 325, 639–644. 24 Solov’ev, V.N., Polikarpov, E.V., Nemukhin, A.V., and Sergeev, G.B. (1999) J. Phys. Chem. A, 103, 6721–6725. 25 Hwang, D.Y. and Mebel, A.M. (2000) J. Phys. Chem. A, 104, 7646–7650. 26 Polikarpov, E.V., Granovsky, A.A., and Nemukhin, A.V. (2001) Mendeleev Commun., 4, 150–151. 27 Hwang, D.Y. and Mebel, A.M. (2000) Chem. Phys. Lett., 331, 526–532. 28 Burkholder, T.R., Andrews, L., and Bartlett, R.J. (1993) J. Phys. Chem., 97, 3500–3503.

85

86

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2 29 Chin, C.H., Mebel, A.M., and Hwang, D.Y. (2003) Chem. Phys. Lett., 375, 670–675. 30 Lequere, A.M., Xu, C., and Manceron, L. (1991) J. Phys. Chem., 95, 3031–3037. 31 Huber, H., McIntosh, D., and Ozin, G.A. (1978) Inorg. Chem., 17, 1472–1476. 32 Kurtikyan, T.S. and Kazaryan, S.G. (1981) Arm. Khim. Zh., 34, 375. 33 Almond, M.J., Downs, A.J., and Perutz, R.N. (1985) Inorg. Chem., 24, 275–281. 34 Mascetti, J. and Tranquille, M. (1985) Surf. Sci., 156, 201–205. 35 Mascetti, J. and Tranquille, M. (1988) J. Phys. Chem., 92, 2177–2184. 36 Mele, F., Russo, N., Toscano, M., and Illas, F. (2001) New Trends in Quantum Systems in Chemistry and Physics. Volume 2: Advanced Problems and Complex Systems, 7, Springer, Dordrecht, The Netherlands, pp. 143–165. 37 Hwang, D.Y. and Mebel, A.M. (2002) Chem. Phys. Lett., 357, 51–58. 38 Papai, I., Schubert, G., Hannachi, Y., and Mascetti, J. (2002) J. Phys. Chem. A, 106, 9551–9557. 39 Papai, I., Mascetti, J., and Fournier, R. (1997) J. Phys. Chem. A, 101, 4465–4471. 40 Zhou, M.F. and Andrews, L. (1999) J. Phys. Chem. A, 103, 2066–2075. 41 Hwang, D.Y. and Mebel, A.M. (2002) J. Chem. Phys. A, 116, 5633–5642. 42 Papai, I., Hannachi, Y., Gwizdala, S., and Mascetti, J. (2002) J. Phys. Chem. A, 106, 4181–4186. 43 Zhou, M.F., Liang, B.Y., and Andrews, L. (1999) J. Phys. Chem. A, 103, 2013–2023. 44 Pantazis, D.A., Tsipis, A.C., and Tsipis, C.A. (2004) Coll. Czech. Chem. Commun., 69, 13–33. 45 Jiang, L., Teng, Y.L., and Xu, Q. (2007) J. Phys. Chem. A, 111, 7793–7799. 46 Galan, F., Fouassier, M., Tranquille, M., Mascetti, J., and Papai, I. (1997) J. Phys. Chem. A, 101, 2626–2633. 47 Mebel, A.M. and Hwang, D.Y. (2000) J. Phys. Chem. A, 104, 11622–11627. 48 Hannachi, Y., Mascetti, J., Stirling, A., and Papai, I. (2003) J. Phys. Chem. A, 107, 6708–6713.

49 Dobrogorskaya, Y., Mascetti, J., Papai, I., and Hannachi, Y. (2005) J. Phys. Chem. A, 109, 7932–7937. 50 Chertihin, G.V. and Andrews, L. (1995) J. Am. Chem. Soc., 117, 1595–1602. 51 Mascetti, J., Galan, F., and Papai, I. (1999) Coord. Chem. Rev., 192, 557–576. 52 Caballol, R., Sanchez Marcos, E., and Barthelat, J.C. (1987) J. Phys. Chem., 91, 1328–1333. 53 Shabatina, T.I., Mascetti, J., Ogden, J.S., and Sergeev, G.B. (2007) Russ. Chem. Rev., 76, 1123–1137. 54 Zhang, L.N., Wang, X.F., Chen, M.H., and Qin, Q.Z. (2000) Chem. Phys., 254, 231–238. 55 Liang, B.Y. and Andrews, L. (2002) J. Phys. Chem. A, 106, 4042–4053. 56 Liu, X.P., Gong, J.K., Collins, A.W., Grove, L.J., and Seyler, J.W. (2001) Appl. Organomet. Chem., 15, 95–98. 57 Chen, M.H., Wang, X.F., Zhang, L.N., and Qin, Q.Z. (2000) J. Phys. Chem. A, 104, 7010–7015. 58 Wang, X.F., Chen, M.H., Zhang, L.N., and Qin, Q.Z. (2000) J. Phys. Chem. A, 104, 758–764. 59 Liang, B.Y. and Andrews, L. (2002) J. Phys. Chem. A, 106, 595–602. 60 Jiang, L. and Xu, Q. (2007) J. Phys. Chem. A, 111, 3519–3525. 61 Tague, T.J., Andrews, L., and Hunt, R.D. (1993) J. Phys. Chem. A, 97, 10920–10924. 62 Beck, W., Raab, K., Nagel, U., and Steimann, M. (1982) Angew. Chem., Int. Ed. Engl., 21, 526–527. 63 Castro-Rodriguez, I., Nakai, H., Zakharov, L.N., Rheingold, A.L., and Meyer, K. (2004) Science, 305, 1757–1759. 64 Chang, C.C., Liao, M.C., Chang, T.H., Peng, S.M., and Lee, G.H. (2005) Angew. Chem., Int. Ed., 44, 7418–7420. 65 Gao, G., Li, F., Xu, L., Liu, X., and Yang, Y. (2008) J. Am. Chem. Soc., 130, 10838–10839. 66 Szczepankiewicz, S.H., Ippolito, C.M., Santora, B.P., Van de Ven, T.J., Ippolito, G.A., Fronckowiak, L., Wiatrowski, F., Power, T., and Kozik, M. (1998) Inorg. Chem., 37, 4344–4352. 67 Contreras, L., Paneque, M., Sellin, M., Carmona, E., Perez, P.J., Gutierrez-

References

68

69

70 71

72 73

74

75

76 77

78

79

80 81

82

83

84

Puebla, E., Monge, A., and Ruiz, C. (2005) New J. Chem., 29, 109–115. Lee, C.H., Laitar, D.S., Mueller, P., and Sadighi, J.P. (2007) J. Am. Chem. Soc., 129, 13802–13803. Gibson, D.H., Ding, Y., Andino, J.G., Mashuta, M.S., and Richardson, J.F. (1998) Organometallics, 17, 5178–5183. Chi, Y., Lan, J.W., Peng, S.M., and Lee, G.H. (2001) J. Cluster Sci., 12, 421–432. Wang, T.F., Hwu, C.C., and Wen, Y.S. (2004) J. Organomet. Chem., 689, 411–418. Jeoung, J.H. and Dobbek, H. (2007) Science, 318, 1461–1464. Ohnishi, T., Seino, H., Hidai, M., and Mizobe, Y. (2005) J. Organomet. Chem., 690, 1140–1146. Srivastava, R., Bennur, T.H., and Srinivas, D. (2005) J. Mol. Cat. A – Chemistry, 226, 199–205. Castro-Rodriguez, I. and Meyer, K. (2005) J. Am. Chem. Soc., 127, 11242–11243. Castro-Rodriguez, I. and Meyer, K. (2006) Chem. Commun., 1353–1368. Fox, A.R., Bart, S.C., Meyer, K., and Cummins, C.C. (2008) Nature, 455, 341–349. Hirano, M., Akita, M., Tani, K., Kumagai, K., Kasuga, N., Fukuoka, A., and Komiya, S. (1997) Organometallics, 16, 4206–4213. Deglmann, P., Ember, E., Hofmann, P., Pitter, S., and Walter, O. (2007) Chem. Eur. J., 13, 2864–2879. Matsubara, T., and Hirao, K. (2001) Organometallics, 20, 5759–5768. Aresta, M., Tommasi, I., Dileo, C., Dibenedetto, A., Narracci, M., Ziolkowski, J., and Jezierski, A. (2001) Can. J. Chem., 79, 570–577. Gong, J.K., Wright, C.A., Thorn, M., McCauley, K., McGill, J.W., Sutterer, A., Hinze, S.M., and Prince, R.B. (1998) Adv. Chem. Conv. Mitig. Carbon Dioxide, 114, 491–494. Konno, H., Kobayashi, A., Sakamoto, K., Fagalde, F., Katz, N., Saitoh, H., and Ishitani, O. (2000) Inorg. Chim. Acta, 299, 155–163. Ohnishi, Y.Y., Nakao, Y., Sato, H., and Sakaki, S. (2006) Organometallics, 25, 3352–3363.

85 Aresta, M., Tommasi, I., Dibenedetto, A., Fouassier, M., and Mascetti, J. (2002) Inorg. Chim. Acta, 330, 63–71. 86 Louie, J. (2005) Curr. Org. Chem., 9, 605–623. 87 Sakakura, T., Choi, J.C., and Yasuda, H. (2007) Chem. Rev., 107, 2365–2387. 88 Sakakura, T. and Kohno, K. (2009) Chem. Commun., 1312–1330. 89 Hoberg, H. and Schaefer, D.J. (1982) J. Organomet. Chem., 236, C28–C30. 90 Hoberg, H., Peres, Y., Kruger, C., and Tsay, Y.H. (1987) Angew. Chem., Int. Ed. Engl., 26, 771–773. 91 Dedieu, A. and Ingold, F. (1989) Angew. Chem., Int. Ed. Engl., 28, 1694–1695. 92 Sakaki, S., Mine, K., Taguchi, D., and Arai, T. (1995) Bull. Chem. Soc. Jpn, 68, 1873–1882. 93 Schubert, G. and Papai, I. (2003) J. Am. Chem. Soc., 125, 14847–14858. 94 Papai, I., Schubert, G., Mayer, I., Besenyei, G., and Aresta, M. (2004) Organometallics, 23, 5252–5259. 95 Graham, D.C., Mitchell, C., Bruce, M.I., Metha, G.F., Bowie, J.H., and Buntine, M.A. (2007) Organometallics, 26, 6784–6792. 96 Aresta, M., Pastore, C., Giannoccaro, P., Kovacs, G., Dibenedetto, A., and Papai, I. (2007) Chem. Eur. J., 13, 9028–9034. 97 Li, J., Jia, G.C., and Lin, Z.Y. (2008) Organometallics, 27, 3892–3900. 98 Laitar, D.S., Mueller, P., and Sadighi, J.P. (2005) J. Am. Chem. Soc., 127, 17196–17197. 99 Zhao, H., Lin, Z., and Marder, T.B. (2006) J. Am. Chem. Soc., 128, 15637–15643. 100 Matsuo, T. and Kawaguchi, H. (2006) J. Am. Chem. Soc., 128, 12362–12363. 101 Sadique, A.R., Brennessel, W.W., and Holland, P.L. (2008) Inorg. Chem., 47, 784–786. 102 Evans, W.J., Seibel, C.A., and Ziller, J.W. (1998) Inorg. Chem., 37, 770–776. 103 Davies, N.W., Frey, A.S.P., Gardiner, M.G., and Wang, J. (2006) Chem. Commun., 4853–4855. 104 Allen, O.R., Dalgarno, S.J., and Field, L.D. (2008) Organometallics, 27, 3328–3330.

87

88

4 Carbon Dioxide Coordination Chemistry and Reactivity of Coordinated CO2 105 Summerscales, O.T., Frey, A.S.P., Geoffrey, F., Clocke, N., and Hitchcock, P.B. (2009) Chem. Commun., 198–200. 106 Choi, J.C. and Sakakura, T. (2003) J. Am. Chem. Soc., 125, 7762–7763. 107 Shi, F., Zhang, Q.H., Ma, Y.B., He, Y.D., and Deng, Y.Q. (2005) J. Am. Chem. Soc., 127, 4182–4183. 108 Takaya, J. and Iwasawa, N. (2008) J. Am. Chem. Soc., 130, 15254–15255.

109 Aresta, M., Quaranta, E., Tommasi, I., Giannoccaro, P., and Ciccarese, A. (1995) Gaz. Chim. Ital., 125, 509–538. 110 Schenck, S., Notni, J., Kohn, U., Wermann, K., and Anders, E. (2006) Dalton Trans., 4191–4206. 111 Aresta, M., Tommasi, I., Quaranta, E., Fragale, C., Mascetti, J., Tranquille, M., Galan, F., and Fouassier, M. (1996) Inorg. Chem., 35, 4254–4260.

89

5 Main Group Element- and Transition Metal-Promoted Carboxylation of Organic Substrates (Alkanes, Alkenes, Alkynes, Aromatics, and Others) Thomas Zevaco and Eckhard Dinjus

5.1 Introduction

Carbon dioxide (CO2) is considered today as one of the “bad guys of the chemical scene,” due mainly to its greenhouse gas character and the connected global warming. As outlined in the course of this book, many ways have been proposed from different communities to control the volume of CO2 in the atmosphere; some of these might have concrete applications, whereas the leftovers fall into the “wishful thinking” category. It is therefore a remarkable incentive for “synthetic” chemists to have a share in solving this global problem and to ﬁnd new promising synthesis routes involving CO2, even though it is clear that chemical ways will still remain a small part of a more comprehensive solution. Carbon dioxide is, from many points of view, an ideal C1-synthon for organic synthesis because of its benignity, low cost, and sheer endless availability as a renewable resource. Therefore, the development of new catalytic methods to activate the otherwise thermodynamically stable CO2 and the resultant formation of new C–C bonds represents a challenge for both academia and industry. From an “atom efﬁciency” point of view, the C–C coupling of CO2 to an organic substrate to form in a few steps valuable products containing the intact COO moiety (carboxylic acids and carbonates) constitutes a good example for a sustainable “green” chemistry. In this chapter attention will be focused on the chemical transformations of CO2 to yield carboxylic acids and related molecules (e.g., cyclic esters), while “natural” carboxylations reactions such as those dealing with RuBisCO (ribulose-1,5 bis(phosphate) carboxylase/oxidase) and its genetic engineering or the carboxylation of phenol via a phenol phosphate intermediate by the phenol-carboxylase enzyme, will be left aside.

Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

90

5 Main Group Element- and Transition Metal-Promoted Carboxylation

5.2 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis 5.2.1 Kolbe–Schmitt Synthesis: Generalities

One of the most reported examples of an industrial synthesis using CO2 as a reagent is the synthesis of aromatic hydroxy-carboxylic acids from the parent alkali metal aryloxides, amino-phenoxides and, to a lesser extent, alkali salts of hydroxypyridines [1]. The products obtained (e.g., 2-hydroxy-benzoic acid, 4-hydroxybenzoic acid, 3-hydroxy-2-naphthoic acid 6-hydroxy-2-naphthoic acid) and the metal salts thereof can be used as starting compounds for pharmaceuticals (aspirin), antibacterial agents and color-developing agents for pressure-sensitive recording sheets and heat-sensitive recording paper [2]. Furthermore, some of these molecules ﬁnd an application as additives for lubricating oils, as building blocks in polyester derivatives, and as specialty liquid crystals [3]. This synthesis, which is also called the Kolbe–Schmitt synthesis, after its inventors, is possibly the oldest organic synthesis using CO2. Kolbe reported, in 1860, the successful synthesis of 2-hydroxy-benzoic acid (also called salicylic acid due to its initial extraction as methyl salicylate from the bark of willow trees; genus: Salix linnaeus) by heating a mixture of phenol and sodium under an atmosphere of CO2 [4]. This solid-state reaction is usually carried out by contacting dry CO2 with alkali metal aryloxides, or by rapidly passing CO2 through the melted aryloxides. The thus-obtained sodium salicylate is then dissolved in water, the carboxylic acid precipitating upon acidiﬁcation with sulfuric acid. The reaction is quite selective with sodium and lithium phenoxides, in general generating small amounts of 4-hydroxy benzoic acid as well as traces of 2-hydroxy-isophthalic acid and 4-hydroxyisophthalic acid as side products. Using the same general procedure, Kolbe also produced p-cresotic acid and o-thymotic acids from p-cresol and o-thymol, respectively [5]. The carboxylation occurs more easily when the aromatic compounds are activated by electron-donating substituents, as in the case of aminophenols or diand trihydroxyaromatic compounds. In order to optimize the general yield of this solid-state reaction (less than 50%, due partly o the formation of disodium salicylate and phenol in equivalent amounts), Schmitt investigated the use of a higher CO2-pressure and reported, in 1885, on greatly improved yields (above 80%) [6]. A typical reaction was performed in an autoclave at 403 K with dry sodium phenoxide “stirred” for several hours under a pressure of CO2 ranging from 8 to 13 MPa, to eventually yield (almost quantitatively) the pure sodium salicylate. One of the shortcomings of this method was that the reaction mixture actually is a waxy solid or a highly viscous liquid with a limited contact surface between alkali metal phenoxide and CO2. The Kolbe– Schmitt reaction has also been reported to take place in solution or suspension, which allows a better stirring of the reaction mixture and thus a better contact between the alkali phenoxide and CO2 [7]. Hence, many solvents have been investigated in this role: polar media such as phenol [8], ketones [9], dimethylsulfoxide

5.2 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis

(DMSO) [10], N-methylcaprolactam, tetrahydrofuran (THF), and tetramethylurea [11] each provided fair yields of the desired acids. Using an apolar medium such as gasoline and xylene affords suspensions, with the medium acting then more as a dispersant [12]; the yields are in those cases slightly better than in the polar media. Although this variant of the Kolbe–Schmitt synthesis can be practically performed as a continuous process, it was found that solvating the alkali metal aryloxides did not provide any signiﬁcant improvement in the yields, and that the solvent-free Kolbe–Schmitt procedure was still unbeaten in term of high yields. This behavior was understood some years later following a computer-aided molecular design [13], which showed that the solvents, although capable of dissolving the alkali phenoxide and improving the reaction’s rate constants, would also allow the reverse reaction (decarboxylation) to take place because the ﬁnal product was retained in the reactive phase. However, this is not the case in the classical solventfree Kolbe–Schmitt reaction, as the ﬁnal step of the reaction mechanism is irreversible. A quite elegant optimization of the Kolbe–Schmitt synthesis was patented by Marassé in 1893 [14]. His modiﬁcation consisted of reacting, at high temperature, a mixture of free phenol and anhydrous potassium, rubidium, or cesium carbonate under CO2-pressure to give as the main product the related salt of the salicylic acid. Interestingly, running the reaction with the cheaper sodium, magnesium and calcium carbonates afforded disappointingly low yields. Interestingly, this reaction has been newly rediscovered by Yamagushi et al. [15], who reported on the use of supercritical carbon dioxide (scCO2) and various alkali and alkaline earth metal carbonates as promoters of the reaction. K2CO3 and KHCO3 displayed the best catalytic activity for salicylic acid synthesis (36% and 17%, respectively), thus conﬁrming the earlier ﬁndings of Marassé. Yamagushi et al. also investigated Lewis acids such as SiO2 and activated ZrO2 (SO2− 4 ), as well as basic oxides (CaO, MgO), and showed these compounds to be inadequate for such reactions in scCO2. In the industrial process used today to manufacture salicylic acid, dry sodium phenoxide obtained from phenol and soda is contacted with CO2 under 0.5 MPa and at temperatures around 373 K. After absorption of approximately one molar equivalent of CO2, the temperature is raised and held at 423–433 K for several hours to fulﬁll the reaction [16]. The ﬁnal “technical-grade” salicylic acid is obtained after successive puriﬁcation steps and, eventually, upon acidiﬁcation with sulfuric or hydrochloric acid (Figure 5.1). 5.2.2 Reaction Parameters and Mechanistic Studies of the Kolbe–Schmitt Synthesis

Speaking more generally of the selectivity of the Kolbe-Schmitt reaction, the applicability of the synthetic method is narrowed by the presence in aryloxides of numerous potential sites for C–C coupling. The reaction of an alkali phenoxide with one molecule of CO2 can, in theory, generate three isomers: 2-hydroxybenzoic; 3-hydroxy-benzoic acid, the formation of which is not expected due to the activating effect of the hydroxyl group (ortho- and para-directing group [17]); and

91

92

5 Main Group Element- and Transition Metal-Promoted Carboxylation Pure phenol

50% NaOH

Water CO2 Phenol Water 1

Mixer

2

Decolorizing agent

3 Filtration

Treatment

Mill autoclave 1. Phenoxide preparation 2. Carboxylation 3. Dilution

60%H2SO4

Precipitation

Figure 5.1

+

OM

Separation

Drier

Technical grade salicylic acid

Typical industrial synthesis of salicylic acid [7].

CO2 0.5-1 MPa 393-413 K

OH -

+ HX +

OH CO2H

CO2 M

- MX

M = Li, Na Scheme 5.1

Formation of salicylic acid from alkali phenoxides and CO2 under pressure.

4-hydroxy-benzoic acid, which is actually also isolated and ﬁnds applications mostly in the polymer industry. In order to understand the general mechanism of the Kolbe–Schmitt reaction and to ﬁne-tune the reaction parameters, many experimental and spectroscopic investigations have been undertaken. In general, these have been tedious, due mainly to the moisture-sensitivity of the starting products and the easy contamination by side products such as inorganic alkali carbonates and hydrogen-carbonates. However, a careful study of the experimental data has led to some general trends. For instance, the nature of the ﬁnal products depends heavily on the alkali cations used in the starting compounds; sodium and lithium phenoxides reacting under similar experimental conditions yield the related salicylates as major products [18] (Scheme 5.1), whereas potassium, rubidium, and cesium phenoxides yield mixtures of 2-hydroxy-benzoic acid and 4-hydroxy-benzoic acid [1] (Scheme 5.2). As a rule of thumb, the yield of p-hydroxybenzoic acid generally increases with the increasing ionic radius of the alkali metal. Both, temperature and CO2pressure were also reported to be paramount in the selectivity of the carboxylation;

5.2 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis O M+

CO2 0.5-1 MPa

OH

-

+

CO2 M

CO2H

+ HX +

+ > 473 K -

CO2 M+

OH

OH

OH

93

- MX CO2H

M = K, Rb Cs

Scheme 5.2 Formation of hydroxy benzoic acids from alkali (K, Rb, Cs) phenoxides and CO2.

operating at lower temperatures produces 2-hydroxy-benzoic acids, whereas 4-hydroxy-benzoic acids need a higher temperature for their formation. Hence, running the carboxylation reaction at a higher temperature will yield 4-hydroxybenzoic acid, even from sodium phenoxide. The same trend was identiﬁed when reacting potassium phenoxide under one atmosphere of CO2, namely that a higher yield of the para product was delivered with increasing temperature. Interestingly, the same reaction when performed under pressure afforded another pattern, with the yield of 4-hydroxy-benzoic acid decreasing with increasing temperature. The same outcome was observed for the carboxylation of potassium and cesium phenoxides at temperatures in the range of 423 to 533 K, at constant pressure [19]. However, this situation was explained by the impossibility in a closed system to distil the phenol formed in situ, so as to displace the equilibrium towards the alkali salts of the hydroxy-carboxylic acids. Among the numerous reaction mechanisms proposed previously [20–22], the most plausible have involved the formation of an alkali metal phenoxide–CO2 complex at an early stage of the reaction. Sodium and potassium phenoxide–CO2 compounds can be easily formed at room temperature after some hours under 0.1 MPa CO2 [11]. At a higher temperature (423 K) and high CO2-pressure, the sodium phenoxide–CO2 complex intermediate undergoes an intramolecular rearrangement to produce the sodium salt of salicylic acid. In comparison, heating such intermediates under normal pressure mostly results in a decarboxylation of the compound and the liberation of phenol. Despite numerous spectroscopic and theoretical studies, the exact structure of the active species (the alkali metal phenoxide–CO2 complex) formed in the ﬁrst stage of the Kolbe–Schmitt reaction remains the subject of much debate. Some of the structures proposed, which range from pure alkali phenyl carbonates to “CO2-solvated alkali phenoxides”, are shown in Figure 5.2 for the case of potassium phenoxide. Early studies postulated the straightforward formation of alkali phenyl carbonate 1 (PhOCO2M, with M = K or Na) as the reactive intermediate [4, 5]. Major discrepancies in the in situmeasured infrared (IR) spectra, such as no characteristic carbonate νas(C=O) absorption band around 1750 cm−1, were the main reasons for the search for more suitable structures. On the basis of further IR studies and differential scanning calorimetry (DSC) analyses [11], it appeared that the alkali metal phenoxide–CO2 complex, when reacted under Kolbe–Schmitt conditions, underwent an irreversible phase transition and evolved to produce at least two further intermediates. In

94

5 Main Group Element- and Transition Metal-Promoted Carboxylation O

K O

O C

O

C

M

O

O

OCO2K 2

101.236

1.407 1.410

1

K

O

O

3

2 1.413

4

1.597

1 5

6

8 1.397

2.594

2.624

K

C O 3 Figure 5.2

13 1.185 12

7* 1.270 9

16 15 14 1.186

1.201

11

2.775

2.772

1.198

Proposed structures for the PhOK–CO2 complex in the Kolbe–Schmitt synthesis.

complementing other studies, this investigation rules out a direct carboxylation at the aromatic ring. Based on the results of these studies, the alkali metal phenoxide–CO2 complex deﬁnitely plays a key role in the formation of the hydroxy benzoic acids, and an understanding of its particular structure would elucidate the ortho/para selectivity of the reaction. Structure 2 is a promising candidate able to explain the para and ortho reaction pathways. This intermediate exhibits weak interactions between π-electrons of the aromatic ring and the carbon atom of CO2, whereas CO2 is activated by the cation and the oxygen-alkali ion bond is weakened. This structure was proposed on the basis of IR and nuclear magnetic resonance (NMR) data [22], and was assumed to further react via C-alkylation at the ortho position and proton transfer to the phenolic oxygen. The existence of structure 2 was later reinforced by an X-ray diffraction (XRD) structural study of the solvent-free sodium phenoxide [11]. This crystal structure displayed as its main feature polymeric chains of Na2O2 four-membered rings generated by the stacking of PhO−Na+ pairs. The structure showed signiﬁcant short distances between the cation of one unit and the aromatic ring of the next unit (average 3.05(8) Å), which spoke in favor of potential weak interactions between an electron-rich aromatic ring and the partly positively charged carbon atom of CO2. Such interaction with the bulky phenyl group led to a remarkable low-coordinated sodium ion suitable to coordinate and activate CO2, but this postulated PhONa–CO2 π-complex could not be conﬁrmed experimentally. Structure 3, as proposed by Kosugi et al., displayed a complex transient carbonate species formed via concurrent interactions between the phenoxide ion, CO2, and the potassium cation [23]. The structure of intermediate 3, as well as the proposed mechanism, were rather speculative, however. Swimming against the mainstream, Kosugi proposed on the basis of NMR experiments with 13C-labeled CO2 and MOPAC/PM3 calculations, that the reaction proceeded by direct carboxylation of

5.2 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis

the aromatic ring. Consequently, the PhONa*CO2 complex was deemed not to be the key intermediate in this process; rather, its formation should be considered more as a side reaction competing with the formation of either salicylic acid or p-hydroxybenzoic acid. Whilst the studies focused mainly on the carboxylation of potassium phenoxide, an exact mechanism conﬁrming their hypothesis was not presented. Considering the experimental difﬁculties and the relative inconsistency of some experimental data, computer-aided molecular design (CAMD) and molecular modeling have attracted ever-growing interest to evaluate the structures of these intermediates and to better understand their activation mechanism(s) in the Kolbe–Schmitt synthesis [24, 25]. The comprehensive studies of Markovic et al. have shed new light on the structure of the intermediate(s), and may actually clarify many experimental ﬁndings. The research group ﬁrst investigated the role and structure of the sodium phenoxide−CO2 complex 4 (Figure 5.2) in the formation of salicylic acid [24a] and, subsequently, also the inﬂuence of the alkali ion on the ortho/para selectivity of the Kolbe–Schmitt synthesis [24b]. Their calculations conﬁrmed that the reactions would most likely proceed in the ortho and para positions, with the noticeable exception of lithium phenoxide which yields only salicylic acid. In fact, the yield of the para-substituted product was found to increase as the ionic radius of the alkali metal used was increased, and that the high ortho selectivity of the lithium and sodium phenoxides was a direct consequence of their short ionic radii. The group of Markovic eventually examined the structure–reactivity relationship of the potassium phenoxide−CO2 complex [24c], with special attention being paid to the last reaction step of the para route of the Kolbe−Schmitt reaction [24d] (Scheme 5.3). These studies, which employed density functional theory (DFT) methods (B3LYP/LANL2DZ/Gaussian 98) proposed that the reactions of all alkali metal phenoxides with CO2 followed a similar ground mechanism that comprised three intermediates and three transition states. In step 1, CO2 must ﬁrst be activated by an alkali metal phenoxide. In the case of the sodium phenoxide [24a], CO2 can only attack at the polarized O–Na bond to form a PhONa/CO2 complex as the ﬁrst intermediate (structure 4). The calculation deﬁnitely rules out a direct C–C bond formation at the aromatic ring. PhONa + CO2 → B → TS1 (o- and p-) → C → TS2 → D → TS3 → Sodium salicylate

(5.1)

Scheme 5.3 Structure of the ﬁrst ortho- and para-intermediates formed in the Kolbe–Schmitt synthesis, as proposed by Markovic et al. [24c, d].

95

5 Main Group Element- and Transition Metal-Promoted Carboxylation 50 40 30

1.446

3

1.382

2

4

1.435

1.453

1

5

4

1.313

1.386

O

1.441 1.308 O 1.462 1

6

3

1.448

2 2.134

2.259

20

2.191

7 O 1.1281.246O

Na

10

2.126

Na 2.187

pTS2-K

PhO-K+CO2 0

pTS1-K

–10

oTS1-K

pC-K oC-K

O

C

O C O M O oC-M O M O O C

M

O

H

pD-K oD-K

oTS2-K

H C O

H

–20 B-K

–30 –40

oTS3-K pTS3-K

1.384

2.120

7 O 1.222 1.246 O

E (kcal mol–1)

96

O

O M O

O M C O H

pC-M

oE-K

oD-M

O

pE-K

pD-M

–50 Reaction coordinate Figure 5.3 Calculated energy diagram for the formation of ortho- and para-hydroxyl benzoic acid from potassium phenoxide (also shown are the transition states p-TS-Na and o-TS-Na for the sodium derivatives) [24c, d].

In step 2 of the reaction, the electrophilic carbon atom of CO2 attacks the ring at the ortho position. The reaction proceeds via the transition states oTS1 and TS2 (Figure 5.3), with formation of the intermediates C and D. The ﬁnal product, sodium salicylate, is obtained from intermediate D through a proton shift from the ring to the phenoxide ion via the transition state TS3. It can be shown from the ﬁgure that intermediates C and D can easily undergo the reverse reactions, thus conﬁrming the role of the CO2-pressure to shift the equilibrium towards the ﬁnal products of the reaction. Interestingly, studies dealing with the potassium phenoxide−CO2 complex [24d, e] and the nature of the intermediates responsible for the ortho and para routes could partly explain the result obtained by Kosugi and coworkers in their experiment with 13C-labeled CO2 [23]. Markovic et al. completed their Kolbe−Schmitt models with a microsolvated model, and distinguished two starting geometries for the potassium phenoxide−CO2 complex: (i) a structure with no supplementary coordinated CO2 molecules (atmospheric pressure); and (ii) a structure with one to two CO2 molecules coordinated at the potassium atom (maximum of two due to computing limits) (Figure 5.4) [24d]. The latter microsolvated structure represents the potassium phenoxide− CO2 complex under Kolbe−Schmitt experimental conditions (high CO2-pressures). Carbon dioxide can then be seen as a solvent stabilizing the KOPh–CO2 complex.

5.2 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis 10 10 1.406 1.410

3

2

4

1.412 1.635

1 5

6

1.233 7* 1.266

8

1.396

9

1.410

3

2

4

1.413 1.597

1 5

2.535

1.407

1.397

1.236 1.270

7*

K

12

8

6

2.594 2.624

K 2.564

13

9

16 15 14

1.185

1.201

11

2.775

2.772

1.186 1.198

Figure 5.4 Calculated structures of a potassium phenoxide–CO2 “nonsolvated” and “solvated” with two CO2 molecules [24d].

It appears that, under the conditions of the Kolbe–Schmitt reaction (high pressure and temperature), the labeled potassium phenoxide–*CO2 complex of Kosugi undergoes a scrambling of labeled and unlabeled CO2 molecules: KOPh-13CO2 ⋅2 ( 12 CO2 ) → TS1⋅2 ( 12 CO2 ) → TS1⋅( 13 CO2 )⋅( 12 CO2 ) → KOPh12CO2K ⋅( 13 CO2 ) ⋅ ( 12 CO2 )

(5.2)

As the ortho and para intermediates formed within the KS synthesis are also CO2solvated, a constant exchange of labeled and unlabeled CO2 molecules will occur such that the amounts of 13C-labeled products of the Kolbe–Schmitt reaction are constantly leveled. The results obtained by Markovic et al. were conﬁrmed, using quantum mechanical calculations, by Stanescu et al. [13, 25], with another software package (Jaguar 4.2.) and a slightly different data set B3LYP/CEP-31+G(d). The energy proﬁles calculated for the reaction path followed the same pattern. Stanescu took this model as basis to understand the inﬂuence of different “solvents” on the course of the Kolbe–Schmitt reaction, and was able to conﬁrm the experimental trends. Markovic and coworkers extended these DFT calculations (B3LYP/LANL2DZ) so as to understand the mechanism of the carboxylation of 2-naphthoxide which yields, at high temperatures, mixtures of 3-hydroxy-1naphthoic acid [26a] as a main product, as well as 2-hydroxy- and 6-hydroxy-2naphthoic acids [26b]. 5.2.3 Recent Applications of the Kolbe–Schmitt Carboxylation: Synthesis of 1,3-Dialkylimidazolium-2-Carboxylates

Tommasi et al. recently reported the carboxylation of 1,3-dialkylimidazolium-2carboxylates, which can be seen as a remarkable variant of the Kolbe–Schmitt synthesis [27]. In this case, it was shown that 1,3-dialkylimidazolium-2carboxylates could be synthesized from 1,3-dialkylimidazolium chlorides and CO2 via a Kolbe–Schmitt-type reaction (actually, this was more of a Marassé variant, run in solution). The starting compounds were carboxylated in anhydrous dimethylformamide (DMF) under approximately 5 MPa CO2, at temperatures ranging from 353 to 408 K and with Na2CO3/CO2 as a “catalyst,” according to Scheme 5.4.

97

98

5 Main Group Element- and Transition Metal-Promoted Carboxylation

P(CO2)=5 MPa 2

1

R

R

Cl 1

+ Na2CO3

2

1

Δ DMF

R

R

+ NaHCO3

C O

O

2

R =R = CH3 1

2

R = CH3, R = n-Bu Scheme 5.4 Formation of 2-dialkyl-imidazolium-carboxylates following the Kolbe–Schmitt synthetic procedure [27].

The yields of the reaction were very high, as was the selectivity of the C–C coupling at carbon 2 (although traces of 4- and 5-carboxylate isomers were identiﬁed at higher temperatures). The synthesis of such 2-imidazolium carboxylates is a topic of growing interest, owing to their use as potential ionic liquids (ILs) with tunable characteristics, and their ability to carboxylate C–H acidic compounds [28]. The results showed that a direct carboxylation could be carried out at relatively moderate temperatures (353–383 K), in good yield, and without the formation of signiﬁcant amounts of side products (Scheme 5.4). For instance, running the reaction at 383 K for 36 h delivered 1,3-dimethylimidazolium-2-carboxylate in high yield (92%) and selectivity (91%), with the 4-carboxylate isomer as a minor side product (9%). The presence of other substituents caused the formation of other isomers, as shown by the case of the 1-butyl-3-methylimidazolium salt which, under the same conditions, produced a mixture of 1-butyl,3-methylimidazolium-2carboxylate (82%), 1-butyl-3-methylimidazolium-4-carboxylate (10%), and 1-butyl-3methylimidazolium-5-carboxylate (8%) [27]. Incidentally, another synthesis route of 1,3-dimethylimidazolium-2-carboxylate was reported by Holbrey et al., using 1-methyl-imidazole as substrate and dimethyl carbonate as a “carboxylating” agent [29a]. This synthesis has been extensively investigated, using DFT calculations (B3PW91), by Voutchkova et al. [29b]. Both, the Kolbe–Schmitt reaction and the dialkyl carbonate route may be beneﬁcial, notably when taking into account the former syntheses used to produce these carboxylato derivatives. Prior to the study of Tommasi et al., the formation of imidazolium carboxylates via highly reactive N-heterocyclic carbenes (NHC) had been reported, although this rather “challenging” synthesis could not be extended to an industrial scale. The synthesis of 1,3-diisopropyl-4,5-dimethyl-imidazolium2-carboxylate from 1,3-diisopropyl-4,5-dimethylimidazol-2-ylidene and CO2 was reported by Kuhn et al. [30]. For this, the reaction of the imidazol-2-ylidene with CO2 provided imidazolium-2-carboxylates, the chemical behavior of which was quite similar to that of “standard” carboxylic acids. This compound could be protonated with HCl or HBF4, and/or alkylated with [Et3O]BF4. Likewise, the cationic imidazolium-2-carboxy chloride could be isolated from the reaction of imidazolium-2-carboxylate with thionyl chloride. A few years later, Louie and coworkers [31] focused on the synthesis and reactivity of 1,3-dimesitylimidazolium-2-carboxylate from the

5.2 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis

carboxylation of the related carbene with CO2. Here, the imidazolium carboxylates were found capable of exchanging their carboxylate groups with less-stable imidazolium carboxylates (transcarboxylation), thus opening new routes for mild carboxylation. A related synthetic method was also reported by Lu and coworkers, which was based on the reaction of 1,3-bis(2,6-diisopropylphenyl) imidazolium salts obtained via deprotonation with KOtBu under an atmosphere of CO2, to afford 1,3bis(2,6-diisopropylphenyl)imidazolium-2-carboxylate in good yield [28e]. This method of activating CO2 was subsequently used in elegant fashion for catalytic coupling reactions involving CO2 and terminal epoxides to yield cyclic carbonates. 5.2.4 Carboxylation of C–H-Acidic Compounds

Those species capable of activating CO2 by forming either a reactive carboxylate precursor or a stable molecule with a loosely bound carboxylate group, represent promising candidates to serve as mild catalysts in the so-called trans- or remote carboxylation reactions of C–H acidic compounds (e.g., nitromethane, acetone, acetophenone, cyclohexanone, and phenylacetonitrile). Among the substrates used, some have attracted interest as possible starting compounds for carboxylates of pharmaceutical interest [32]. However, very few reports detailing the carboxylation of these compounds (which are also known as “active methylene” or “active hydrogen” compounds) with CO2 have been produced. The ﬁrst catalysts to be investigated were, almost logically, potassium and sodium phenoxides [33]. Thus, Bottaccio et al. [33b] postulated the formation of phenol and a highly active enolate species stabilized by an appropriate aprotic solvent (Scheme 5.5). In a successive step, the enolate reacts with CO2 to produce the alkaline salt of the carboxylic acid. Later, Mori and coworkers reported on the carboxylation in DMF of cyclohexanone and nitromethane via different monosubstituted potassium phenoxides [33d, f].

M= Li, Na, K OM

R= linear, branched alkyl or alkoxy group up to C20 n=1 to 5

Rn

substrate / phenoxide molar ratio: 1/1 to 1/4 O

H R' CH R'' "active hydrogen" substrate ketones, esters, nitroparaffins, nitriles

0.1 MPa CO2

C

OM

R' CH R''

313K > 323 K > 333K aprotic solvent: THF, NEt3, EtOAc,...

Scheme 5.5 Carboxylation of C–H-active substrates with alkali phenoxides and CO2 [33].

99

100

5 Main Group Element- and Transition Metal-Promoted Carboxylation

Walther et al. improved the NaOPh/CO2 system by tuning the role of the solvent, such that the carboxylation of acetone was carried out selectively by sodium phenoxide with N-methyl-caprolactam under 0.1 MPa CO2 to afford 3-ketoglutaric acid in 65% yield [34]. The same authors reported the conversion of acetylnaphthalene (acetonaphthone) to acetyl-naphthalene carboxylic acid with 84% yield, and of cyclohexanone to cyclohexane-2,6-dicarboxylic acid in 56% yield. More recently, Kunert et al. have investigated bulky substituted zinc phenoxides in the carboxylation of benzophenone. In this case, zinc(II) 2,6-dimethylphenoxide and zinc(II) 2,6-di-tert-butyl-phenoxide [35] produced the best results at 313 K in DMSO (46 and 37% yields, respectively). The study was complemented with DFT studies that shed some light on the structure/13C NMR of the zinc phenoxide–CO2 complex, but without delving into certain mechanistic investigations. Other alkoxides which have also been extensively investigated include the lanthanoid alkoxides (La, Sm, Yb) that were used by Inoue et al. in the carboxylation of ﬂuorene, ketones, and esters [36]. The system was based on the formation in situ of the actual CO2carrier by reacting the alkoxide with phenyl isocyanate under a CO2 atmosphere. The postulated CO2-transfer agent displayed an iminodicarboxylate structure (Scheme 5.6) that was capable of carboxylating C–H acidic compounds. Most other reported catalytic systems rely on the use of strong nucleophiles such as 1,3-diphenyl urea (DPU), diazabicyclo[5.4.0]undec-7-ene (DBU) [37], diphenyl carbodiimide (DPC) [38] and anilides [39] to activate CO2 and acidic substrates. R2 Ln–OR1

N

R2N=C=O

+

C

Ln

OR1

O 1

149.0

CO2 R2 R3

H

Ln

O

161.2

R2 N H

C O

OR1

Ln

C

N

C

OR1

O

O

148.9

2 R3

O C O 3

153.2

178.4

H3O+ H

O

C O

Scheme 5.6

R3 171.8

Carboxylation of C–H-active substrates with lanthanoid alkoxides and CO2 [36].

5.2 Formation of Aromatic Carboxylic Acids: The Kolbe–Schmitt Synthesis H N

COCH3 + CO32–

– N

COCH3 + CO2

N

COCH3 + AMCH

– N

N C

COCH3 + HCO3–

COCH3

O

O–

H N

COCH3

+

AMC–COO–

C O

O–

Scheme 5.7 Proposed mechanism for the CO2-transfer performed by acetanilides in the carboxylation of C–H-active substrates [39].

For example, DBU was investigated in the carboxylation of cyclohexanone to obtain 2-oxo-cyclohexane-carboxylic acid. Relying on kinetic data, a mechanism was tentatively proposed in which DBU abstracted a proton from the substrate, thus enabling the generated carbanion to react with CO2 [37]. A direct activation in the form of a “DBU·CO2” complex (as in the case of alkali metal phenoxides with CO2) was excluded. By comparison, Chiba et al. conducted extensive studies in DMSO of several acetanilides and formanilides in combination with K2CO3 to carboxylate ﬂuorene into 9-ﬂuorene-carboxylic acid (yield 27%) [39] (Scheme 5.7). This catalytic system was also capable of carboxylating indene to 1-indenecarboxylic acid in good yields (54%), whereas acetone and cyclohexanone each delivered low yields of the carboxylated products (14%). Chiba suggested, as an active catalytic species, the existence of an “acetanilide–CO2” complex, which was a type of carbamate formed from the reaction of the deprotonated acetanilide with CO2. Interestingly DPU, which was also reported in this study, provided the best result in terms of the carboxylation of ﬂuorene (47% yield). The ﬁnal system reported by Chiba involved the use of DPC and alkali carbonate under a CO2 atmosphere [38]. In this case, similar reaction trends could be found, and indene could be carboxylated in good yield (54%), whereas ﬂuorene afforded 9-ﬂuorene-carboxylic acid in only 35% yield. It was possible to enhance the carboxylation of ﬂuorene either by using cesium carbonate as a cocatalyst (up to 72% yields, otherwise 38% with K2CO3), or by using an activated carbodiimide such as bis(para-methyl-phenyl) carbodiimide (up to 43% yield). The mechanism proposed was based on an activation of the diaryl carbodiimide by the carbonate ion so as to produce a stronger nucleophile capable of activating CO2 and transferring the so-formed carboxylate group to the C–H-acidic substrate. A similar mechanism was also conceivable when using DPU instead of DPC as the “catalyst”. Unfortunately, these synthetic methodologies delivered relatively low yields of carboxylated products, or required an excess of catalyst for quantitative conversion of the substrates.

101

102

5 Main Group Element- and Transition Metal-Promoted Carboxylation O CH3

N

N

″Bu + R-H + NaI

R

C

+

O– Na+

CH3 N

″Bu

I–

C O

N

O–

Scheme 5.8 Formation of sodium carboxylates with 2-dialkyl-imidazolium-carboxylate as the CO2-carrier [40].

In comparison, the 1,3-dialkylimidazolium-2-carboxylate isolated by Tommasi et al. [27, 40] was revealed to be a more versatile catalyst that allowed the synthesis of benzoylacetic acid from benzophenone and CO2 in good yield and under mild conditions (isolated yield 81%). The presence of tetraﬂuoroborate- or tetraphenylborate sodium salts in the reaction was essential, as this allowed the formation of the related 1,3-dialkylimidazolium tetraﬂuoroborate or tetraphenylborate and the concomitant quantitative trans-carboxylation to sodium benzoylacetate. Likewise, compounds such as acetone, cyclohexanone, and phenylacetonitrile could also be converted with this system to afford the corresponding carboxylate salts (methyl α-cyanophenylacetate) (Scheme 5.8). Following the same general procedure, acetone was carboxylated, being simultaneously the cosolvent and reagent. In this case, the reaction was run for 70 h at room temperature to selectively afford sodium 3-oxo-butanoate in high yield (77%). The carboxylation of cyclohexanone to afford selectively the 2-oxo-cyclohexane-1-carboxylate product (yield 62%) and the phenylacetonitrile carboxylation yielding sodium α-cyanophenylacetate (60% yield) were carried out in CH3CN as the solvent. The way in which 1,3-dialkylimidazolium-2-carboxylate activates the substrate and transfers the carboxylate moiety remains unclear, but it has been suggested that the determinant step in the trans-carboxylation is the regeneration of a free, active carbene capable of reacting with the organic substrate to generate a carbanion. This species would then react easily with CO2 to build the new carboxylate. By considering the optimized synthesis of the 1,3-dialkylimidazolium-2-carboxylates, and their promising reactivity as CO2-transfer reagents in remote carboxylations, Tommasi suggested that both might be integrated in a single process. It should be noted that the dimethyl-imidazolium chloride depicted in Scheme 5.9 plays the role of the “catalyst” in the synthetic process.

5.3 Reactive Organometallic Derivatives in the Synthesis of Carboxylic Acids 5.3.1 Generalities

From an historical point of view, the production of carboxylic acids from CO2 and reactive organometallic derivatives dates back almost as far as the Kolbe–Schmitt

5.3 Reactive Organometallic Derivatives in the Synthesis of Carboxylic Acids

R-C(O)ONa

CH3-N

N-CH3 Cl-

PCO2 = 5 MPa + Na2CO3

metathesis with NaCl

R-H + NaBPh4

CH3-N –O

N-CH3

NaHCO3 + NaCl

O

R=CH3O, PhC(O)CH2 Scheme 5.9 Integrated catalytic process using 2-dialkyl-imidazolium-carboxylates as the CO2-carrier [40].

synthesis. This century-old story was initiated in 1901 by Victor Grignard and his academic teacher Philippe Barbier, at the University of Lyon in France [41]. The so-called “Grignard reagents” and their related organolithium counterparts [42] are strong nucleophiles that are known to react swiftly with CO2 to directly form valuable carboxylic acids and related products. The chemistry of organomagnesium compounds is characterized by a polarized carbon–magnesium bond, which provides the carbon atom with a negative partial charge. The resultant carbanions will then react with the partially positive charged carbon of CO2, while the metallic cation will simultaneously activate one of the oxygen atoms of CO2. Taking into consideration the difference in electronegativity between carbon and other elements based on the Allred–Rochow scale (Li: 1.53, Mg: 1.27, Zn: 0.84, Sn: 0.78, B:0.49, H: 0.30, C: 0.00), it becomes clear that the carbon–magnesium bond will be less polarized than the carbon–lithium bond. This, in turn, explains the lower reactivity of the Grignard reagents, which generally is coupled with a better selectivity of the reaction. Consequently, organolithium and organomagnesium compounds will both generally be more reactive towards electrophiles than will organozinc, organoboron or organotin compounds. Organocopper [43] and organomanganese [44] reagents are also known to react with CO2, albeit to a lesser extent. Despite the high reactivity and (eventually) good yields, the most frequent drawback associated with these standard reactions is that they are stoichiometric in nature. In other words, unwanted side products may be formed that might represent a supplementary problem for the environment. In the same fashion, the restricted functional group compatibility of Grignard reagents and other highly reactive organometallic reagents may eventually limit their application in the synthesis of complex carboxylic acids. The development of new methods by which to conduct versatile carboxylations with CO2 has long been the subject of many research projects. One method chosen to overcome this reactivity problem was to generate a transient, milder organome-

103

104

5 Main Group Element- and Transition Metal-Promoted Carboxylation

tallic reagent, the reactivity with CO2 of which could be easily tuned with another metallic system. As a consequence, when several research groups focused their interest on such bimetallic carboxylating “catalytic” systems, four signiﬁcant systems have been reported as promising alternatives. 5.3.2 Bimetallic Catalytic Systems 5.3.2.1 Pd/Sn Systems Shi et al. reported one of the ﬁrst examples of bimetallic catalytic systems that allowed the insertion of CO2 into the rather unreactive tin–carbon bond [45]. The concept behind this system was to exploit, in the same system, the ability of a transition metal to catalyze crosscoupling reactions and CO2 activation. For instance, tributyl(allyll)tin does not react with CO2 in solution even under highpressure. To run the same reaction in the presence of zero-valent palladium species (Pd(PPh3)4 or Pd(PBu3)4) will quantitatively afford carboxylates 2 (90%) and 3 (10%) (Scheme 5.10), although the reactivity of the system is limited to allylstannanes. The tentative mechanism postulates a transfer of the allylic moiety to the palladium(0) via oxidative coupling, followed by an insertion of CO2 into the Pd–C bond (as noted previously [46]). The so-formed allyl-carboxylato ligand can then be transferred back to the tin atom to liberate in the next step a tin carboxylate. The active Pd(0) species is then regenerated via reductive elimination of the carboxylate- and organotin groups The isomerization observed in the reaction was not clearly elucidated, however. This Pd/Sn catalytic system was subsequently extended by Franks et al., who described the palladium-catalyzed “carboxylative coupling” of organostannanes with organic halides under CO2 pressure to produce allylic esters in good yields [47]. The mechanism proposed was similar to the former scheme with a supplementary trans-esteriﬁcation step at a palladium center. Transfer of a second allylic moiety to a reactive palladium(I)-allowed formation of the ester after reductive elimination and regeneration of the active palladium(0) species. As noted earlier for the carboxylation of tributyl(allyll)tin, this bimetallic system is subject to signiﬁcant ligand scrambling, which somehow diminishes the synthetic utility of the method (Scheme 5.11). Johansson et al. [48] were interested in improving this bimetallic system, and proposed the existence of PCP-pincer ligands (2,6-bis[(di-t-butylphosphino)methyl]

Scheme 5.10 system [45].

Formation of tin carboxylates with a dual allylstannanes/Pd(PPh3)4 catalytic

5.3 Reactive Organometallic Derivatives in the Synthesis of Carboxylic Acids O CH3 +

Bu3sn 1a

2b

O 3a

CO2(5 MPa)

Cl

CH3 O

MLn THF, 343 K, 24 h

+

O

O

CH3 O +

O 3c

CH3

3b

O 3d

CH3 + Bu3SnCl

Catalst

%3a

%3b

%3c

%3d

Pd(PPh3)4

20

30

21

29

Pd(PPh3)2Cl2

24

25

21

20

Scheme 5.11 Carboxylative coupling of organostannanes with organic halides, showing a scrambling of the ligands during the carboxylation/esteriﬁcation step [47].

Figure 5.5 Molecular structure of 2,6-bis[(di-t-butylphosphino)methyl]phenyl Pd-Me, and of its CO2-insertion product [48].

phenyl-) as new alternatives to the palladium(0) precursors of Shi and Franks. For instance, the (PCP)Pd–Me complex displaying an activated Pd–C bond could undergo a CO2-insertion to form, quantitatively, (PCP)Pd–OAc (Figure 5.5). Both, the starting compound and the carboxylated product could be structurally characterized. The catalytic capabilities of this system towards crosscoupling reactions have been successfully evaluated with dimethylzinc to yield Zn(OAc)2, with the methyl complex being regenerated using ZnMe2. Incidentally, Dong et al. recently studied the reactivity of more common palladium(0) and palladium(II) precursors towards transmetallation/carboxylation tandem reactions with aryl-zinc bromide reagents. [Pd(PCy3)2], whether used as is

105

106

5 Main Group Element- and Transition Metal-Promoted Carboxylation PPh2 O Pd O

CF3

PPh2 13 Bu3Sn

O Bu3Sn

(a) O

CF3 PPh2 Pd PPh2 16

O Bu3Sn

CO2

O slow

Bu3Sn

(c)

(b)

fast

PPh2 O Pd O PPh2 17

Scheme 5.12 Proposed mechanism for the carboxylation/transmetallation performed by the (PCP)Pd species of Johanson and coworkers [50].

or generated in situ from Pd(OAc)2, catalyzed very efﬁciently the formation of benzoic acid from the related organozinc reagents [49]. Subsequent to the ﬁrst PCP-pincer study, Johansson used NMR spectroscopy to investigate the reactivity of the PCP–Pd complexes in transmetallation reactions with organotin derivatives. Somewhat surprisingly, only those palladium complexes with electron-withdrawing groups and less sterically demanding substituents could react efﬁciently with tributylallyltin [50]. Compared to systems using Pd(0), this system was shown to be more selective (affording only one deﬁnite regioisomer) and to function at a lower CO2 pressure (0.4 instead of 3.3 MPa), even though the reaction delivered a lower yield of carboxylate. The proposed mechanism involved the formation of an active PCP–Pd(II) allyl complex which would then undergo a rapid insertion of CO2 into the Pd–C to produce a PCP–butenoate species (Scheme 5.12). This complex could then undergo transmetallation with allyl stannane to give, once again, the allyl complex and thus completing the catalytic cycle.

5.3 Reactive Organometallic Derivatives in the Synthesis of Carboxylic Acids

Scheme 5.13 Rhodium-catalyzed carboxylation of alkylboronic esters under mild conditions [51].

5.3.2.2 Rh/B and Cu/B In 2006, Ukai et al. proposed an interesting alternative with a rhodium(I)-catalyzed carboxylation of aryl- and alkenylboronic esters proceeding under mild conditions, and leaving ancillary reactive functional groups such as carbonyl- and cyano unreacted [51] (Scheme 5.13). Considering that organoboronic esters are easily available, and that various functional groups are tolerated, this reaction appeared to be particularly useful for the preparation of functionalized arylcarboxylic acids, such as benzoic and cinnamic acid derivatives. However, whilst this catalytic reaction provided an efﬁcient route to a variety of derivatives, the corresponding alkylboronic esters could not be carboxylated with the Rh(I) system under the investigated conditions. This lack of reactivity was most likely due to a coordination of these functional groups to the rhodium metal, and the related inactivation of the catalyst. It was found that arylboronic esters containing a bromide would generate complex mixtures of products, thus limiting to some extend the synthetic utility of this aproach. For these reasons, the same research group extended their studies to other metallic systems, and reported the details of an exciting boronic ester/copper(I) salt/CsF catalytic system which could also very efﬁciently promote this type of carboxylation reaction with a far wider functionalgroup compatibility [52]. The best results were obtained with a system using copper(I) iodide in combination with a bis-oxazoline as the stabilizing ligand; this gave substituted benzoic acid in high yields and purities (up to 95% for paramethoxy-benzoic acid), while the presence of cesium ﬂuoride was paramount to activate the boronic esters (Scheme 5.14). The investigated system was shown to tolerate an even wider range of ligands and copper(I) precursors. Hence, by using tetramethylethylenediamine (TMEDA), 1,4-diazabicyclo[2.2.2]octane (DABCO), bipyridine or diphenylphosphino-propane as ligands, and copper(I) acetate or chloride as precursors, signiﬁcant amounts of benzoic acid were afforded from the neopentyl glycolato ester of phenylboronic acid. The major improvement compared to the Rh(I) catalysts was the ability of the Cu(I)/RB(OR′)2 system to also carboxylate alkenyl-boronic esters in very good yield, under similar conditions. Interestingly, the system functioned better under ligandless conditions, producing the corresponding unsaturated carboxylic acids in good yields. A reaction mechanism was tentatively proposed by Takaya et al. which involved, in a ﬁrst step, the formation of aryl- or alkenyl-copper(I) species formed by

107

108

5 Main Group Element- and Transition Metal-Promoted Carboxylation

Scheme 5.14

Optimized copper-catalyzed carboxylation of alkylboronic esters [52].

Scheme 5.15 esters [52].

Proposed mechanism for the copper-catalyzed carboxylation of alkylboronic

transmetallation of the in situ-generated cesium ﬂuoroborates with the Cu(I) salt [52]. The highly reactive organo-copper(I) intermediate underwent “as usual” a CO2 insertion into the M–C bond to produce a copper(I)-carboxylate (Scheme 5.15). Via ligand exchange with the ﬂuoroborate and subsequent protonation, the carboxylic acid could ﬁnally be isolated. 5.3.2.3 Ni/Zn The most often-reported metallic systems used in the formation of carboxylic acids are most certainly those systems that involve a zero-valent nickel species as the active intermediate. Ochiai et al. reported on a bimetallic catalytic system which allowed the synthesis of various saturated carboxylic acid in good yields, under very mild conditions [53] (0.1 MPa CO2, 4–8 h reaction time, temperatures ranging from room temperature to 323 K). The catalytic system was based on the use of organozinc reagents as carbon nucleophiles, which could be selectively carboxylated in the presence of Ni(acac)2 as the main catalyst. As an example, a hexylzinc iodide-lithium chloride complex quickly reacted in THF under 0.1 MPa CO2 in the presence of Ni(acac)2 and tricyclohexyl-phosphane as catalyst, to produce heptanoic acid in 70% yield (Scheme 5.16). In their catalytic cycle, the authors proposed the formation in situ of a zero-valent nickel species by the reduction of Ni(acac)2 with the organozinc reagent. Since the pioneering studies of Aresta et al., which detailed the isolation of the ﬁrst nickel– CO2 complex, such zero-valent species have been known easily to activate CO2 [54]. Transfer of the so-formed Ni(II)-carboxylate occurred following a transmetallation

5.3 Reactive Organometallic Derivatives in the Synthesis of Carboxylic Acids R–Znl•LiCl + CO2

cat. Ni(acac)2 cat. P(c-C6H11)3 rt

R–CO2H R = alkyl

(0.1 MPa) Scheme 5.16 Ni-catalyzed carboxylation of organozinc iodide reagents under mild conditions [53].

H Nickel Catalyst X

+ CO2 + R2Zn

X

(0.1 MPa)

CO2H *

* *

R

H OMe PPh2

10 Examples 90–96% ee

MeO-MOP Scheme 5.17 Dual Ni(0)/ZnR2 catalytic system with separate carboxylation and transmetallation steps [55f].

with the organozinc iodide reagent, after which a reductive elimination step regenerated the zero-valent nickel species and afforded concurrently the desired zinc carboxylate; this, upon protonolysis, yielded the ﬁnal product. The role of lithium chloride as “cocatalyst” however was not elucidated, the authors speculating that LiCl activated the organozinc iodide by changing its aggregation state and creating a more-reactive monomer species. The synthetic method was successfully extended to other phosphines and organozinc precursors, and showed great promise. Other research groups have also concentrated their investigations on the versatile Ni/Zn system. For example, Yeung et al. studied the reactivity of [Ni(PCy3)2] under crosscoupling/carboxylation conditions, and were able to isolate (in high yields) benzoic acid from the related arylzinc bromide (respectively iodide) reagents [49]. The positive inﬂuence of lithium chloride on the course of such catalysis has been demonstrated in the same way. Later, the research group of Mori concentrated on a slightly different approach, notably the multistep syntheses which make effective use of the dual Ni(0)/ZnR2 system so as to yield complex organic molecules (e.g., erythrocarine [55a], tamoxifen [55b] and chaetomellic acid A anhydride [55c]). The main difference in these studies was that the carboxylation and transmetallation steps did not necessarily occur at the same active center. In other words, within one catalytic cycle, the Ni/Zn catalyst system could perform more than one C–C coupling, namely a carboxylation at location a via an active Ni(0) intermediate, and an extra C–C bond formation at location b via the organozinc species. This could, of course, be complemented by concurrent regioselective and stereoselective ring-closing reactions [55d–f] (Scheme 5.17).

109

110

5 Main Group Element- and Transition Metal-Promoted Carboxylation

R CO2 + 2 R

Ni(COD)2 / PR3 CH3CN / THF 1 MPa CO2 , 353 K

R

O

O

4 Scheme 5.18

Ni(0)-catalyzed formation of substituted 2-pyrones.

More generally speaking, the oxidative cyclization of CO2 and unsaturated hydrocarbons using zero-valent nickel species to produce reactive oxanickelacycles (cyclic nickel-carboxylates) has been the subject of massive research interest [56– 58], initiated by the pioneering studies of Hoberg and coworkers [59]. This was due to the fact that these versatile intermediates could be used in many ways for the synthesis of various functionalized carboxylic acids and ester derivatives (a second organometallic counterpart, e.g., ZnR2, is not necessarily required). Both, alkynes and alkenes react generally in a similar manner with CO2 in the presence of the Ni(0) species. However, the formation of unsaturated nickelacycles via the coupling of alkynes with CO2 was found to generate, after the ﬁnal work-up, a broader range of synthetically useful α,β-unsaturated carboxylic acids and cyclic esters (Scheme 5.18). The bulk of the relevant literature has referred to, on the one hand, the formation of 2-pyrones and poly-pyrones from alkynes using Ni(1,5cyclo-octadiene)2/phosphine catalytic systems [56, 57], on the other hand with the formation of saturated/unsaturated carboxylic acids via electrochemical methods [58]. The formation of 2-pyrones from CO2 and alkynes was ﬁrst described by Inoue, who used a combination of Ni(cod)2 and chelating phosphines as catalysts [56a]. Although initially, the yields were low (even under drastic reaction conditions), over the years the synthesis was investigated and optimized by two research groups. These studies evolved into two synthetic procedures capable of producing 2-pyrones in high yields and selectivities. The ﬁrst method involved the use of THF–acetonitrile as solvent, in combination with basic phosphines with small cone angles, such as P(C2H5)3, whereas the second method used trimethyl phosphine and scCO2 both as a single solvent and, of course, as the substrate [56g]. Kishimoto et al. recently reported an improved catalytic system which functioned in scCO2 [56f], where the synthesis can be performed from 4 up to 15 MPa with a combination of Ni(cod)2 and P(C4H9)3 or P(C8H17)3 to afford a wide range of substituted pyrones in high yields (up to 98%) and selectivities. The second well-documented research area that involved zero-valent nickelcatalyzed carboxylation was initiated by Dunach et al. during the 1990s [58a–c]. These studies showed that electrochemically generated Ni(0)- and Mg(II)-based species represented good alternatives to the catalysts used in standard synthetic procedures. The formation of unsaturated carboxylic acids was proposed as a catalytic reaction initiated by Ni(0), although the exact mechanism and the obligatory

5.3 Reactive Organometallic Derivatives in the Synthesis of Carboxylic Acids

presence of magnesium ions (from a sacriﬁcial magnesium anode) in order to perform the reaction have still not been completely elucidated. Following another lead, Silvestri and coworkers focused on more reactive unsaturated substrates, and investigated the electrocarboxylation reactions of benzyl halides through redox catalysis [58d]. The electrochemical reduction of organic halides represents a very convenient way in which to generate carbanions capable of reacting with CO2, and also to build the related carboxylic acids, as further demonstrated in the electrocarboxylation of chloroacetonitrile to cyanoacetic acid [58d]. To conclude the details on zero-valent nickel carboxylation catalysts, some recent synthetic approaches worthy of note showed that this area of research still has a rich chemistry. For example, Louie and coworkers reported on the use of Nheterocyclic carbenes (diaryl-imidazolylidene) as new efﬁcient ligands in the Nicatalyzed coupling of various symmetrical di-ynes with CO2 (Scheme 5.19) [60a]. This pyrone synthesis illustrated that substrates containing steric demanding groups at the terminal positions did not undergo cyclization, whereas “asymmetrical” di-ynes containing bulky groups and methyl terminal groups afforded the desired pyrones, with high regioselectivity [60c]. Another example of promising research is the efﬁcient electrochemical dicarboxylations of aryl-acetylenes with CO2, using an uncomplicated bimetallic redox couple as the catalytic system. In this case, metallic nickel was used as the cathode and aluminum as the anode, to generate in situ “carboxylation-active” nickel species (Scheme 5.20) [61]. Under optimal conditions ((n-bu4NBr) / DMF as the supporting electrolyte, room temperature, CO2 pressure 2–3 MPa, electricity at 4 F mol−1), the electrochemical

R3 R1

R3

R2

R3

5 mol % Ni(COD)2 10 mol % IPr

R1

0.1 MPa CO2, 333 K, 2h

R2

N N

O O

(1)

R3

IPr Scheme 5.19 Ni-catalyzed carboxylative cyclization, yielding pyrones under mild conditions [60a].

Scheme 5.20 Ni-catalyzed electrochemical carboxylation of aryl-acetylenes without stabilizing ligands [61].

111

112

5 Main Group Element- and Transition Metal-Promoted Carboxylation

route could afford the corresponding aryl-maleic anhydrides and 2-arylsuccinic acids in excellent yields (82–94%). Interestingly, besides the synthetic approach of the Ni-catalyzed carboxylation of unsaturated hydrocarbons, two groups have simultaneously investigated, via DFT calculations (DFT B3LYP/LANL2DZ basis set for Ni and 6-31G* basis set for CO2 and alkynes) the potential structures of the nickelacycle intermediates, as well as the probable reaction pathways involved in these reactions. Hence, Li et al. focused on the regioselectivity observed in the Ni(0)-mediated coupling reactions of both terminal and internal alkynes with CO2 [62a]. Graham et al. investigated the inﬂuence of different parameters such as acetylene-substituent and ancillary ligands such as TMEDA and DBU on the potential energy surface of the reaction, in an attempt to rationalize the effects of these parameters on the overall reaction energetics and regioselectivity [62b]. The results gained from the DFT calculations performed on the Ni(0)/acetylenes/CO2 catalytic system correlated with related theoretical studies conducted with the stoichiometric Ni(0)-promoted ethylene/CO2 coupling reaction to yield acrylic acid [63]. Similar to the acetylene/CO2 coupling reaction, the proposed reaction pathway involving ethylene and CO2 ﬁrst proceeds through the side-on-coordination of ethylene to the zero-valent nickel center, followed by a subsequent attack of the CO2 carbon on one of the carbons of the coordinated ethylene to form a nickelacycle. This reaction most likely occurs in a concerted manner, with the concurrent formation of a new carbon–carbon bond (coupling CO2-ethylene) and of a nickel– oxygen bond (activation of CO2 and formation of an oxanickelacycle). The existence of a Ni(bipy)(CO2)(C2H4) complex in solution is unlikely, due to the higher stability of the Ni(bipy)(C2H4) complex. A thorough investigation of the role of the co-ligand showed that the relative stability of the intermediates involving a C–C bond formation and the related activation barrier depended heavily on the nature of the chelating ligands.

5.4 Palladium (0)-Catalyzed Telomerization of Butadiene with CO2: Synthesis of δ-Lactone

The ﬁrst reports describing the reaction of 1,3-butadiene with CO2 to yield cyclic esters were made by Inoue [64] and Musco [65] during the late 1970s. These studies showed that the telomerization of 1,3-butadiene with CO2 afforded a broad range of products with various synthetic uses, including δ- and γ-lactones (1–2), 1,3,7-octatriene (3), carboxylic acids (4 and 5) and, in smaller quantities, openchained octadienyl esters of nonatriene carboxylic acid (6 and 7). The research groups of Behr [66] and Braunstein [67] investigated more in detail the reaction path to the most representative molecule of this product range: the δ-lactone 2-ethylidene-6-hepten-5-olide (1) (Scheme 5.21). All of the efﬁcient catalytic systems reported to date have involved the use of palladium species, together with various ligand architectures such as [Pd(η3-2-MeC3H4)(OAc)], Pd(acac)2 with PnBu3 and PiPr3, and [Pd(PPh3)2(p-benzoquinone)] [68].

5.4 Palladium (0)-Catalyzed Telomerization of Butadiene with CO2: Synthesis of δ-Lactone

Pd(0) O

CO2

C

O

O

1

C O

2

Pd(0)

CO2H

CO2

Pd(0)

CO2H

2 Pd(0)

CO2

3 2

4

+ O

C

5

Pd(0)

O

C

O

O

6

7

Scheme 5.21 Pd(0)-catalyzed telomerization of butadiene and CO2.

Here, one of the favorite starting compounds is Pd(acac)2 which affords, under optimal conditions, signiﬁcant yields of δ-lactone with a very high selectivity. The active Pd(0) catalytic species is actually formed during the reaction, being stabilized via either tri-alkyl- or tri-aryl-phosphines. More speciﬁcally, it seems that phosphine ligands with a high σ-donator character and bulky substituents (such as isopropyl and cyclohexyl) greatly facilitate the catalysis [69]. Although, many research groups have investigated the rules of formation of the zero-valent palladium species, no conclusive mechanism has yet been demonstrated. Among the mechanistic studies describing the formation of δ-lactone, Behr and coworkers showed that the presence of nitrile groups in the reaction medium was paramount to obtain the six-membered ring lactone in high yields (Scheme 5.22) [70]. The existence of a phosphine/nitrile synergy has been proposed to explain the high catalytic activity of this system, with acetonitrile stabilizing the catalytically active zero-valent palladium/phosphine complex via the formation of a Pd(NCR)x(PR3)4 complex. The acetonitrile ligands are more loosely bound to the Pd(0) center than their phosphinic counterparts, and are thus able easily to generate free coordination sites for the butadiene coordination. This so-called “nitrileeffect” was elegantly used by Pitter and coworkers to produce, as result of a

113

114

5 Main Group Element- and Transition Metal-Promoted Carboxylation

CO2 + 2

Pd(acac)2 / PR3 CH3CN 2.5 MPa CO2 , 363 K

O

O

2 Scheme 5.22 Typical “nitrile-assisted” synthesis of the δ-lactone: 2-ethylidene-6-hepten-5-olide.

ligand-design, new hemi-labile phosphino-nitrile ligands of the general formula R2P-(CH2)n-CN (best results with n > 5) [71a]. These P,N-ligands enhanced the versatility of the system, allowing the catalysis to be run in solvents other than acetonitrile, both polar and apolar [71b]. Hence, the use of “CO2-expanded” butadiene (i.e., in solvent-free conditions) afforded a butadiene conversion of up to 95%, with a good lactone selectivity. More importantly, the use of these CO2-expanded liquids allows the process to facilitate the ﬁnal product separation steps, and thus to reduce its environmental impact (i.e., no supplementary solvent needed, no obligatory use of high pressure and temperatures to go supercritical [72]). Based on the particular case on the hemi-labile phosphino-nitrile ligand, a representative general mechanism is summarized in Scheme 5.23. It is suggested, following the related mechanism of the telomerization of butadiene, that the catalytic cycle starts with a phosphino zero-valent palladium species A. The successive coordination and C–C coupling of two butadiene units, with the simultaneous elimination of one ligand (in this case a nitrile ligand), generates a reactive allylic intermediate C. This species then undergoes a CO2-insertion into one Pd–C bond to yield a cyclic metalla-carboxylato species D. The latter compound is also a key intermediate to understand the formation of lactones and open-chained esters. The desired δ-lactone 1 is formed by reductive elimination, thus regenerating the active Pd(0)/Phosphine species A. Interestingly, considering the chiral structure of the δ-lactone, it seems that its enantioselective synthesis is achievable. However, despite many efforts to perform such a catalysis with chiral phosphines or others co-ligands, this has not yet been achieved [73]. In spite of the academic signiﬁcance of this synthesis, the industrial value of δ-lactone as a versatile starting material has remained low because of the few applications reported to date. Dinjus and coworkers [74] described the synthesis of polymers via the photoinitiated polyaddition of δ-lactone with various dithiols which displayed intact lactone fragments within the polymer backbone. However, the complicated synthesis and the formation of rather statistic polymers with hardto-tune characteristics hindered their potential industrial applications. Behr and coworkers have reported on some promising approaches using δlactone as the starting compound for exploitable specialty chemicals. One such strategy involves the three-step synthesis of 2-ethylheptanoic acid, an interesting building block for new alkyd resins (as used in printing inks, baked enamels, and

5.4 Palladium (0)-Catalyzed Telomerization of Butadiene with CO2: Synthesis of δ-Lactone

Pd N

R2P

C B

Pd N

R2P

R2P

C

Pd

C

A C N

O

R2P

CO2

Pd O

O O 1

C

D

N Scheme 5.23 Proposed mechanism for the Pd(0)-catalyzed telomerization of 1,3-butadiene with CO2 (example with hemi-labile phosphino-nitrile ligands) [71].

lubricants) or as a polymer additive (e.g., plasticizers) [75]. In the ﬁrst step, an isomeric mixture of aliphatic carboxylic acids is obtained via a homogeneous hydrogenation with a rhodium(I) catalyst and ring-opening of the δ-lactone. The ﬁnal product 2-ethylheptanoic acid can be isolated in high yields after hydrogenation of the unsaturated bonds, using a commercial available heterogeneous palladium/charcoal catalyst in methanol. In further studies conducted by Brehme et al., a broad variety of reactions such as hydroformylation, hydroaminations and hydrogenation were examined, leading to a wide product range that included diols, dicarboxylic acids and aldehydes of potential industrial interest [76]. The most recent investigations of Behr and coworkers constituted a further step towards high-value chemicals, describing the synthesis in a single step of an unsaturated highly functionalized amino acid, 2-ethylidene-7-morpholinohept-5-enoic acid, obtained from the hydroamination of the δ-lactone with morpholine [77]. The selective 1,4-addition of morpholine to the double bond of the δ-lactone was catalyzed by a platinum catalyst, Pt(cod)Cl2. It is not only a wide range of applications that is important to insure a signiﬁcant industrial up-scaling, but also an ingenious process design allowing easy catalyst retrieval. By relying on catalytic systems that functioned in acetonitrile, Behr et al. tested numerous polyalcohols as extraction media to separate the palladium

115

116

5 Main Group Element- and Transition Metal-Promoted Carboxylation

catalyst from the reaction mixture [78]; subsequently, 1,2,4-butanetriol proved to be the best extraction solvent. The reaction was then up-scaled, the presence of an integrated extraction unit allowed both efﬁcient catalyst recycling and a continuous production of lactone. Pitter et al., by focusing on ligand design, investigated another approach by immobilizing the homogenous catalyst on polystyrene(Merriﬁeld) and silica supports [79]. These supported catalysts could deliver conversions and selectivities similar to the homogeneous systems, but with an easy recycling of the catalyst. Bearing in mind that acetonitrile might be rather problematic for an industrial up-scaling, Behr and coworkers systematically tested different solvents and devised an alternative to the solvent-free phosphino-nitrile ligand system of Pitter et al. by using organic carbonates (linear and cyclic) as the main solvents [78, 80]. Cyclic carbonates (ethylene carbonate, propylene carbonate, butylene carbonate and glycerol carbonate esters) favored much more the formation of δ-lactone than did linear carbonates (dimethyl- and diethyl carbonate). Compared to the results obtained with acetonitrile, the conversion obtained with carbonate solvents was signiﬁcantly improved, with the selectivity of the reaction depending heavily on the size of the substituent in the carbonate solvent.

References 1 Lindsey, A.S. and Jeskey, H. (1957) Chem. Rev., 57 (4), 583. 2 Nishimura, T., Kamashita, T., Wada, M., Nagata, T., Tanabe, Y., Hasegawa, R., and Nakatsuka, M. (1997) US Patent 5647896, Mitsui Toatsu Chemicals. 3 (a) Ueno, R., Masada, Y., and Mori, T. (1991) US Patent 5011984, Ueno Pharmaceutical Co. (b) Ueno, R., Masada, Y., Kuwae, Y., and Kawata, K. (1989) EP 0327221, Ueno Pharmaceutical Co. 4 (a) Kolbe, H. and Lautemann, E. (1860) Justus Liebigs Ann. Chem., 113, 125. (b) Kolbe, H. and Lautemann, E. (1860) Justus Liebigs Ann. Chem., 115, 157. (c) Kolbe, H. (1874) J. Prakt. Chem., 10, 89. 5 Kolbe, H. (1874) US Patent 150867A. 6 Schmitt, R. (1885) J. Prakt. Chem., 31, 397. 7 Boullard, O., Leblanc, H., and Besson, B. (1997) Ullmann’s Encyclopedia of Industrial Chemistry, 5th edn, Wiley-VCH Verlag GmbH, Weinheim, Germany. 8 (a) Dow Chemical Co. (1964) GB Patent 949988. (b) Jansen, G. and Wolff, P. (1983) US Patent 4376867.

9 Fry, D.J. (1950) GB Patent 638196, Ilford Ltd. 10 Furuya, M., Nagatomo, A., and Wada, M. (2002) US Patent 6392090 B1, Mitsui Toatsu Chemicals. 11 Kunert, M., Dinjus, E., Nauck, M., and Sieler, J. (1997) Chem. Ber./Recueil, 130, 1461./ 12 (a) Ueno, R. and Tetsuya, N. (1974) US Patent 3816521, Ueno Pharmaceutical Co. (b) Moseley, J.C., Edyvean, W., and Hartley, J. (1995) GB Patent 738359, Shell Reﬁning & Marketing Co. 13 Stanescu, I. and Achenie, L.E.K. (2006) Chem. Eng. Sci., 61, 6199 and references cited therein. 14 Marassé, S. (1893) GB Patent 189317002. 15 (a) Iijima, T. and Yamaguchi, T. (2007) Tetrahedron Lett., 48, 5309. (b) Iijima, T. and Yamaguchi, T. (2008) Appl. Catal. A: Gen., 345, 12. 16 (a) Hardy, W. L. and Snell, F.D. (1957) Ind. Eng. Chem., 49 (6), 55a. (b) Sakai, T., Ishiguro, T., and Ishihara, T. (1989) US Patent 4814495, Idemitsu Petrochemical Co., Ltd. 17 Vollhardt, K.P.C. and Schore, P.N. (1999) Organic Chemistry, Structure and

References

18

19

20 21 22

23

24

25 26

27 28

Function, 3rd edn, W.H. Freeman & Co., New York. Baine, O., Adamson, G.F., Barton, J.W., Fitch, J.L., Swayampati, D.R., and Jeskey, H. (1954) J. Org. Chem., 19, 510. Wessely, F., Benedikt, K., Benger, H., Friedrich, G., and Prillinger, F. (1950) Monatsh. Chem., 81, 1071. Gilman, H., Arntzen, C.E., and Webb, F.J. (1945) J. Org. Chem., 10, 374. Hales, J.L., Jones, J.L., and Lindsey, A.S. (1954) J. Am. Chem. Soc., 76, 3145. Ayres, D.C. (1966) Carbanions in Synthesis, Oldbourne Press, London, pp. 168–173. (a) Kosugi, Y., Imaoka, Y., Gotoh, F., Rahim, M.A., Matsui, Y., and Sakanishi, K. (2003) Org. Biomolec. Chem., 1, 817. (b) Kosugi, Y., Rahim, M.A., Takahashi, K., Imaoka, Y., and Kitayama, M. (2000) Appl. Organometal. Chem., 14, 841. (c) Kosugi, Y., Takahashi, K., and Imaoka, Y. (1999) J. Chem. Res. (Synopses), 2, 114. (d) Rahim, M.A., Matsui, Y., and Kosugi, Y. (2002) Bull. Chem. Soc. Jpn, 75, 619. (a) Markovic, Z., Engelbrecht, J.P., and Markovic, S. (2002) Z. Naturforsch., 57a, 812. (b) Markovic, Z., Markovic, S., and Begovic, N. (2006) J. Chem. Inf. Model., 46 (5), 1957. (c) Markovic, Z., Markovic, S., Manojlovic, N., and Predojevic-Simovic, J. (2007) J. Chem. Inf. Model., 47 (4), 1520. (d) Markovic, Z. and Markovic, S. (2008) J. Chem. Inf. Model., 48 (1), 143. Stanescu, I., Gupta, R.R., and Achenie, L.E.K. (2006) Mol. Simul., 32, 279. (a) Markovic, Z., Markovic, S., and Durovic, I. (2008) Monatsh. Chem., 139, 329. (b) Markovic, S., Durovic, I., and Markovic, Z. (2008) Monatsh. Chem., 139, 1169. Tommasi, I. and Sorrentino, F. (2006) Tetrahedron Lett., 47, 6453. (a) Aresta, M., Tkatchenko, I., and Tommasi, I. (2003) Ionic Liquids as Green Solvents: Progress and Prospects, vol. 856 (eds R.D. Rogers and K.R. Seddon), ACS Symposium Series, ACS, Washington DC, p. 93.

29

30 31

32

33

34

(b) Scurto, A.M., Brennecke, S.N.V.K., and Aki, J.F. (2002) J. Am. Chem. Soc., 124, 10276. (c) Smiglak, M., Holbrey, J.D., Grifﬁn, S.T., Reichert, W.M., Swatloski, R.P., Katritzky, A.R., Yang, H., Zhang, D., Kirichenko, K., and Rogers, R.D. (2007) Green Chem., 9, 90. (d) Tudose, A., Demonceau, A., and Delaude, L. (2006) J. Organomet. Chem., 691, 5356. (e) Zhou, H., Zhang, W.Z., Liu, C.H., Qu, J.P., and Lu, X.B. (2008) J. Org. Chem., 73, 8039. (a) Holbrey, J.D., Reichert, W.M., Tkatchenko, I., Bouajila, E., Walter, O., Tommasi, I., and Rogers, R.D. (2003) Chem. Commun., 28. (b) Voutchkova, A.M., Feliz, M., Clot, E., Eisenstein, O., and Crabtree, R.H. (2007) J. Am. Chem. Soc., 129, 12834. Kuhn, N., Steimann, M., and Weyers, G. (1999) Z. Naturforsch., 54b, 427. Duong, H.A., Tekavec, T.N., Arif, A.M., and Louie, J. (2004) Chem. Commun., 112. (a) Owen, O.E., Morgan, A.P., Kemp, H.G., Sullivan, J.M., Herrera, M.G., and Cahill, G.F. (1967) J. Clin. Invest., 46, 1589. (b) Hiraide, A., Dohi, S., Suzuki, M., and Shiba, Y. (2002) EP 1188437 (A2), BTG INT Ltd. (c) Kis-Tamas, A., Jurak, F., Vig, Z., Fekete, P., and Kulcsar, J. (1988) US Patent 4740228, Egyt Gyogyszervegyeszeti Gyar. (a) Patmore, E.L., Siegart, W.R., and Chafetz, H. (1972) US Patent 3658874, Texaco Inc. (b) Bottaccio, G. and Chiusoli, G.P. (1966) Chem. Commun., 17, 618. (c) Bottaccio, G., Chiusoli, G.P., Alneri, E., Marchi, M., and Lana, G. (1977) US Patent 4032555, Montedison Fibre S.p.A. (d) Mori, H. and Satake, Y. (1985) Chem. Pharm. Bull., 33, 3469. (e) Mori, H., Okubo, M., Oke, Y., Noguchi, N., Fukuda, M., and Ishihara, M. (1984) Chem. Pharm. Bull., 32, 2200. Walther, D., Ritter, U., Geßler, S., Sieler, J., and Kunert, M. (1994) Z. Anorg. Allg. Chem., 620, 101.

117

118

5 Main Group Element- and Transition Metal-Promoted Carboxylation 35 Kunert, M., Brauer, M., Klobes, O., Görls, H., Dinjus, E., and Anders, E. (2000) Eur. J. Inorg. Chem., 8, 1803. 36 Abe, H. and Inoue, S. (1994) J. Chem. Soc., Chem. Commun., 1197. 37 Mori, H. (1988) Bull. Chem. Soc. Jpn, 61, 435. 38 Chiba, K., Tagaya, H., Karasu, M., Ono, T., Hashimoto, K., and Moriwaki, Y. (1991) Bull. Chem. Soc. Jpn, 64, 966. 39 Chiba, K., Tagaya, H., Karasu, M., Ishizuka, M., and Toshiyuki, S. (1994) Bull. Chem. Soc. Jpn, 67, 452. 40 (a) Tommasi, I. and Sorrentino, F. (2005) Tetrahedron Lett., 46, 2141. (b) Tommasi, I. and Sorrentino, F. (2009) Tetrahedron Lett., 50, 104. 41 (a) Lai, Y.H. (1980) Synthesis, 8, 585. (b) Silverman, G.S. and Rakita, P.E. (eds) (1996) Handbook of Grignard Reagents (Chemical Industries), Marcel Dekker, New York. 42 March, J.L. (1992) in Advanced Organic Chemistry, 4th edn, John Wiley & Sons, Inc., New York, pp. 933–934 and references therein. 43 (a) Cahiez, G., Normant, J.F., and Bernard, D. (1975) J. Organomet. Chem., 94, 463. (b) Fukue, Y., Oi, S., and Inoue, Y. (1994) J. Chem. Soc. Chem. Commun., 2091. (c) Oi, S., Fukue, Y., Nemoto, K., and Inoue, Y. (1996) Macromolecules, 29, 2694. 44 Walther, D., Ritter, U., Kempe, R., Sieler, J., and Undeutsch, B. (1992) Chem. Ber., 125, 1529. 45 Shi, M. and Nicholas, K.M. (1997) J. Am. Chem. Soc., 119, 5057. 46 Hung, T., Jolly, P.W., and Wilke, G. (1980) J. Organomet. Chem., 190, C5. 47 Franks, R.J. and Nicholas, K.M. (2000) Organometallics, 19, 1458. 48 Johansson, R., Jarenmark, M., and Wendt, O.F. (2005) Organometallics, 24, 4500. 49 Yeung, C.S. and Dong, V.M. (2008) J. Am. Chem. Soc., 130, 7826. 50 Johansson, R. and Wendt, O.F. (2007) Dalton Trans., 488. 51 Ukai, K., Aoki, M., Takaya, J., and Iwasawa, N. (2006) J. Am. Chem. Soc., 128, 8706.

52 Takaya, J., Tadami, S., Ukai, K., and Iwasawa, N. (2008) Org. Lett., 10 (13), 2697. 53 Ochiai, H., Jang, M., Hirano, K., Yorimitsu, H., and Oshima, K. (2008) Org. Lett., 10, 2681. 54 (a) Aresta, M., Nobile, C.F., Albano, V.G., Forni, E., and Manassero, M. (1975) J. Chem. Soc. Chem. Commun., 15, 636. (b) Aresta, M., Gobetto, R., Quaranta, E., and Tommasi, I. (1992) Inorg. Chem., 31, 4286. 55 (a) Shimizu, K., Takimoto, M., and Mori, M. (2003) Org. Lett., 5, 2323. (b) Shimizu, K., Takimoto, M., Mori, M., and Sato, Y. (2006) Synlett, 18, 3182. (c) Takimoto, M., Kawamura, M., Mori, M., and Sato, Y. (2005) Synlett, 13, 2019. (d) Takimoto, M. and Mori, M. (2002) J. Am. Chem. Soc., 124, 10008. (e) Takimoto, M., Nakamura, Y., Kimura, K., and Mori, M. (2004) J. Am. Chem. Soc., 126, 5956. (f) Takimoto, M., Shimizu, K., and Mori, M. (2001) Org. Lett., 3, 3345. (g) Shimizu, K., Takimoto, M., Sato, Y., and Mori, M. (2005) Org. Lett., 7, 195. 56 (a) Inoue, Y., Itoh, Y., Kazama, H., and Hashimoto, H. (1980) Bull. Chem. Soc. Jpn, 53, 3329. (b) Burkhart, G. and Hoberg, H. (1982) Angew. Chem., Int. Ed. Engl., 21, 76. (c) Walther, D. and Dinjus, E. (1982) Z. Chem., 22, 228. (d) Hoberg, H. and Schaefer, D. (1984) J. Organomet. Chem., 266, 321. (e) Walther, D., Schönberg, H., Dinjus, E., and Sieler, J. (1987) J. Organomet. Chem., 334, 377. (f) Kishimoto, Y. and Mitani, I. (2005) Synlett, 14, 2141. (g) Jessop, P.G. and Leitner, W. (eds) (1999) Chemical Synthesis Using Supercritical Fluids, Wiley-VCH Verlag GmbH, Weinheim. 57 (a) Tsuda, T., Yasukawa, H., Hokazono, H., and Kitaike, Y. (1995) Macromolecules, 28, 1312. (b) Tsuda, T., Yasukawa, H., and Komori, K. (1995) Macromolecules, 28, 1356. (c) Tsuda, T., Ooi, O., and Maruta, K. (1993) Macromolecules, 26, 4840.

References

58

59

60

61 62

63

64

(d) Tsuda, T. and Maruta, K. (1992) Macromolecules, 25, 6102. (e) Tsuda, T., Kitaike, Y., and Ooi, O. (1993) Macromolecules, 26, 4956. (a) Derien, S., Dunach, E., and Perichon, J. (1991) J. Am. Chem. Soc., 113, 8447. (b) Dunach, E. and Perichon, J. (1988) J. Organomet. Chem., 352, 239. (c) Dunach, E. and Perichon, J. (1990) Synlett, 143. (d) Scialdone, O., Galia, A., Silvestri, G., Amatore, C., Thouin, L., and Verpeaux, J.N. (2006) Chem. Eur. J., 12, 7433. (e) Scialdone, O., Galia, A., Belﬁore, C., Filardo, G., and Silvestri, G. (2004) Ind. Eng. Chem. Res., 43, 5006. (a) Hoberg, H., Schaefer, D., and Oster, B.W. (1984) J. Organomet. Chem., 266, 313. (b) Hoberg, H., Peres, Y., Milchereit, A., and Gross, S. (1988) J. Organomet. Chem., 345, C17. (c) Hoberg, H. and Bärhausen, D. (1989) J. Organomet. Chem., 379, C7. (a) Louie, J., Gibby, J.E., Farnworth, M.V., and Tekavec, T.N. (2002) J. Am. Chem. Soc., 124, 15188. (b) Corrigendum (2004) J. Am. Chem. Soc., 126, 8590. (c) Tekavec, T.N., Arif, A.M., and Louie, J. (2004) Tetrahedron, 60, 7431. Yuan, G.Q., Jiang, H.F., and Lin, C. (2008) Tetrahedron, 64, 5866. (a) Graham, D.C., Bruce, M.I., Metha, G.F., Bowie, J.H., and Buntine, M.A. (2008) J. Organomet. Chem., 693, 2703. (b) Graham, D.C., Mitchell, C., Bruce, M.I., Metha, G.F., Bowie, J.H., and Buntine, M.A. (2007) Organometallics, 26, 6784. (c) Li, J., Jia, G., and Lin, Z. (2008) Organometallics, 27, 3892. (a) Papai, I. and Schubert, G. (2003) J. Am. Chem. Soc., 125, 14847. (b) Papai, I., Schubert, G., Mayer, I., Besenyei, G., and Aresta, M. (2004) Organometallics, 23, 5252. (c) Aresta, M., Pastore, C., Giannoccaro, P., Kovacs, G., Dibenedetto, A., and Papai, I. (2007) Chem. Eur. J., 13, 9028. Sasaki, Y., Inoue, Y., and Hashimoto, H. (1976) J. Chem. Soc. Chem. Commun., 605.

65 Musco, A., Perego, C., and Tartiari, V. (1978) Inorg. Chim. Acta, 28, L147. 66 (a) Behr, A., Juszak, K.D., and Keim, W. (1983) Synthesis, 7, 574. (b) Behr, A. (1988) Carbon Dioxide Activation by Metal Complexes, WileyVCH Verlag GmbH, Weinheim. 67 (a) Braunstein, P., Matt, D., and Nobel, D. (1988) J. Am. Chem. Soc., 110, 3207. (b) Braunstein, P., Matt, D., and Nobel, D. (1988) Chem. Rev., 88, 747. 68 (a) Musco, A. (1980) J. Chem. Soc., Perk. Trans. 1, 693. (b) Behr, A. and Juszak, K.D. (1983) J. Organomet. Chem., 255, 263. (c) Daniels, J.A. (1982) EP 0050445, Imperial Chemical Industries PLC. 69 Leitner, W. and Dinjus, E. (1995) Appl. Organomet. Chem., 9, 43. 70 (a) Behr, A. and Heite, M. (2000) Chem. Eng. Technol., 23, 952. (b) Behr, A. and Heite, M. (2000) Chem. Ing. Tech., 72, 58. 71 (a) Pitter, S., Dinjus, E., and Jung, B. (1996) Z. Naturforsch., Teil B, 51, 934. (b) Pitter, S. and Dinjus, E. (1997) J. Mol. Catal. A: Chemistry, 125, 39. 72 Buchmüller, K., Dahmen, N., Dinjus, E., Neumann, D., Powietzka, B., Pitter, S., and Schön, J. (2003) Green Chem., 5, 218. 73 Storzer, U., Walter, O., Zevaco, T., and Dinjus, E. (2005) Organometallics, 24, 514. 74 Haack, V., Dinjus, E., and Pitter, S. (1998) Angew. Makromol. Chem., 257, 19. 75 Behr, A. and Brehme, V.A. (2002) J. Mol. Catal. A: Chemistry, 187, 69. 76 (a) Behr, A., Brehme, V.A., Ewers, C.L.J., Grüon, H., Kimmel, T., Kuippers, S., and Symietz, I. (2003) Chem. Ing. Tech., 75, 417. (b) Behr, A. and Brehme, V.A. (2002) Adv. Synth. Catal., 344, 525. 77 Behr, A., Henze, G., Johnen, L., and Reyer, S. (2008) J. Mol. Catal. A: Chemistry, 287, 95. 78 (a) Behr, A. and Heite, M. (2000) Chem. Eng. Technol., 23, 11. (b) Behr, A. and Becker, M. (2006) Dalton Trans., 4607.

119

120

5 Main Group Element- and Transition Metal-Promoted Carboxylation 79 (a) Holzhey, N., Pitter, S., and Dinjus, E. (1997) J. Organomet. Chem., 541, 243. (b) Holzhey, N. and Pitter, S. (1999) J. Mol. Catal. A: Chemistry, 146, 25. (c) Holzhey, N., Pitter, S., and Dinjus, E.

(1998) WO 9857745, Forschungszentrum Karlsruhe GmbH. 80 Behr, A., Bahke, P., Klinger, B., and Becker, M. (2007) J. Mol. Catal. A: Chemistry, 267, 149.

121

6 The Chemistry of N–CO2 Bonds: Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas Eugenio Quaranta and Michele Aresta

6.1 Introduction

Carbon dioxide (CO2) can interact easily with several N-nucleophiles. Such interaction results in an activation of the heterocumulene, and has a great synthetic relevance as it is a key step towards the carboxylation or, more generally, the carbonylation of the N-donor substrate and the synthesis of a variety of N-carbonyl compounds. The ﬁxation of CO2 by amines, which has been recognized for a long time, can either take place directly or it can be mediated by metal or nonmetal species to provide a carbamate group, “RR′NCO2”, which is bound either ionic or covalently to an electrophilic center [1–5]. The direct interaction of CO2 with primary or secondary amines can afford carbamic acids or alkylammonium carbamates (Equations 6.1 and 6.2) [1, 2]. In the presence of metals, metal salts (Equation 6.3) [3j, k] or metal-complexes, metal carbamates can be obtained [3, 4]: RR ′NH + CO2 RR ′NCO2H

(6.1)

2RR ′NH + CO2 (RR ′NH2 ) O2CNRR ′

(6.2)

RR ′NH + L + MBPh 4 + CO2 → M (O2CNRR ′ ) + [HL ]BPh 4

(6.3)

where R′ = H, alkyl, L = RR′NH; R = aryl, R′ = H, L = NR 3′′ (R ′′ = alkyl ); M = Li, Na, K Formally, metal or p-block carbamates can be obtained also by insertion of the heterocumulene in the M–N bond of metal- or p-block-amides [4, 5]. All of these compounds, whether alkylammonium, metal and p-block carbamates, are potential carriers or sources of carbamic groups, and can be used in the synthesis of organic carbamates [6] if the transfer of the RR′NCO2 moiety to a suitable organic substrate is accomplished [6, 7]. Moreover, alkylammonium, metal and p-block carbamates, as well as carbamic acid esters, are also potential Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

122

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

precursors of ureas [8] and/or isocyanates [9]. So far, much effort has been expended in attempts to develop new CO2-based synthetic methodologies for these classes of chemicals. The aim of this chapter is to provide an updated account of the research activity in this area. Usually, the traditional routes to the synthesis of carbamic acid esters, isocyanates and ureas start from the toxic compound phosgene [10]. Consequently, the ability to synthesize these compounds from ubiquitous, cheap and safe CO2, as an alternative starting material to phosgene, represents an appealing and highly ecofriendly variation to the conventional protocols which have in the past been based on phosgenation. Moreover, it also provides an attractive answer to the current widespread requests for environmental protection [11].

6.2 Synthesis of Carbamic Acids and Alkylammonium Carbamates

N-Substituted carbamic acids or their anions are believed to play a key role as intermediates in several biological processes, including carboxylations by biotin enzymes and the photosynthetic activation of CO2, as promoted by RubisCO (ribulose 1,5-biphosphate carboxylase) [12]. These species may also be involved in the ﬁxation of CO2 by the amino groups of biopolymers in the solid state [13]. Recently, the formation of carbamate species on surface primary or secondary amino groups of mesoporous graphitic carbon nitride (mpg-C3N4) has been postulated as an intermediate step in the mpg-C3N4-catalyzed oxidation of benzene to phenol, using CO2 as an oxidant in the presence of bases (NaHCO3, triethylamine) [14]. Although carbamic acid, H2NCO2H, has not yet been isolated, its presence has been detected using neutralization-reionization mass spectrometry (NRMS) and, at low temperatures, by infrared (IR) spectroscopy [15]. According to the IR results, it has been proposed that the stability of this compound in the solid phase may be due to intermolecular hydrogen bonding of the zwitterion form, H3N+COO−. The formation of carbamic acid coordinated to Pd(II) through the nitrogen atom has been documented by nuclear magnetic resonance (NMR) spectroscopy (13C resonance at 174.3 ppm), and proposed to be an intermediate step in the mechanism of hydrolysis of urea catalyzed by Pd(II)-complexes [16]. Some properties of H2NCO2H have been the subject of theoretical studies [17]; indeed, calculations have predicted that the syn conformer is more stable than its anti counterpart [17a]. It has been also calculated that gaseous carbamic acid can decompose exothermically into ammonia and carbon dioxide (ΔH° = −26 kJ mol−1) [17b]. Moreover, ab initio calculations have shown that the zwitterion H3N+CO2− is less stable than monomeric acid H2NCO2H and, therefore, the zwitterion form has been proposed as the most probable intermediate for the decomposition of carbamic acid to ammonia and CO2 [17c]. Generally, carbamic acids, RR′NCO2H, are elusive species due to their tendency to decarboxylate, thus giving back CO2 and the amine. Their isolation has been

6.2 Synthesis of Carbamic Acids and Alkylammonium Carbamates

accomplished only recently using either dibenzylamine (Bz2NH) or the Coaminophosphane complex CoCl(NO)2[PhP(OCH2CH2)2NH] [18]. In both cases, X-ray characterization has shown that the relevant carbamic acid molecules are not in the zwitterionic form, but rather are organized in a H-bonded dimeric structure [18, 19]. A feature common to the dimeric moieties of both the isolated acids is the O–H….O distance of 122 pm, which is close to that found in dimeric carboxylic acids. Spectroscopic (IR) and gas-volumetric measurements have demonstrated that [Bz2NC(O)OH]2 forms by the carbonation of neat dibenzylamine through the intermediacy of (Bz2NH2)O2CNBz2. Evidence of the formation of N-substituted carbamic acids has been provided also in solution both for primary and secondary amines [20, 21]. Both NMR and IR analyses have shown that ω-(1-naphthyl)alkylamines in protophilic, highly dipolar, aprotic solvents, such as dimethylsulfoxide (DMSO), dimethylformamide (DMF) and pyridine, are converted quantitatively into the corresponding carbamic acids by reaction with CO2. The formation of mono- and dialkyl-substituted carbamic acids has also been documented in supercritical carbon dioxide (scCO2) by using NMR spectroscopy [22]. The appearance of a low ﬁeld resonance at 8.5–12 ppm in the proton spectrum, and a signal at 158–160 ppm in the 13C spectrum, has been assumed to be diagnostic of the formation of the carbamic acid functionality N–C(O)OH. The formation of carbamic acid species RR′NCO2H in scCO2 has been successfully exploited for the temporary protection of amine function, and has provided a suitable tool for controlling the selectivity of a few catalytic processes carried out in scCO2 as the reaction solvent [22a–c]. The reactivity of some industrially relevant amino-functional silanes, including H2N(CH2)3Si(OMe)3, H2N(CH2)3Si(OEt)3, H2N(CH2)2NH(CH2)3Si(OMe)3 and H2NC(O)NH(CH2)2NH(CH2)3Si(OMe)3, with CO2 has been investigated by Aresta et al. [23]. With the exception of the ureido-derivative, the above-considered compounds promptly reacted with CO2. The kinetics of CO2 uptake showed that, at 295 K, H2N(CH2)3Si(OMe)3 and H2N(CH2)3Si(OEt)3 each reacted with carbon dioxide in a 2 : 1 molar ratio, affording classic ammonium carbamates of formula (RNH3)O2CNHR, while at 273 K dimeric carbamic acids, [RNHC(O)OH]2, were formed [23b]. Conversely, H2N(CH2)2NH(CH2)3Si(OMe)3 reacted at 297 K with CO2 to afford a zwitterionic intramolecular, six-membered cyclic ammonium carbamate of formula (MeO)3Si(CH2)3NH2+CH2CH2NHCO2−, which involved CO2 uptake by the diamine with a 1 : 1 molar ratio [23b]. This chemistry has been more recently revisited by Mehdi, Corriu and coworkers, and exploited to develop a new synthetic approach towards ordered and highly amine-functionalized organosilicas [24]. With respect to carbamic acids, alkylammonium carbamates, (RR′NH2)O2CNRR′ (R′ = H, alkyl), are relatively more stable compounds, and have been recognized for a long time. Under anhydrous conditions, while aliphatic tertiary amines do not absorb CO2 [25], aliphatic primary and secondary amines easily react with CO2 to yield alkylammonium carbamates which, in a few cases, have been fully characterized in the solid state using X-ray diffraction (XRD) [2, 18, 26]. For this,

123

124

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

(RNH3)O2CNHR

RNHCO2H + RNH2 -1

1725 cm

-1

1678 cm , upon

RNHCO2H

CO2 labeling (b)

[RNHCO2H]2 -1

1685 cm

-1

(c)

(R = benzyl)

1643 cm , upon

Scheme 6.1

13

RNH2 + CO2

2 RNHCO2H

(a)

(R = benzyl)

13

CO2 labeling

Chemical stability of alkylammonium carbamates.

R2NH + CO2 HR2N+CO2- + R2NH

HR2N+CO2(R2NH2)O2CNR2

(a) (b)

Scheme 6.2 Mechanism of formation of alkylammonium carbamates (according to Danckwerts [28a]).

strictly anhydrous conditions are required in order to avoid the competitive formation of hydrogenocarbonato- or carbonato-species. In solution, under an inert gas atmosphere (N2), alkylammonium carbamates (RNH3)O2CNHR (R = Bz, allyl, Cy) exhibit modest stability, even at ambient temperature. They readily form carbamic acids RNHCO2H, which can either decarboxylate or undergo self-association with the formation of dimers (Scheme 6.1) [27]. Scheme 6.2 illustrates the pathway proposed by Danckwerts for the conversion of amines and CO2 into alkylammonium carbamates [28a]. The mechanism implies the intermediate formation of a zwitterion species which, by deprotonation, converts into the carbamate product. This reaction pathway has been doubted by Crooks, who proposed a termolecular single-step mechanism [28b]. Borowiak and coworkers [28c] have modeled the reaction of dimethylamine with CO2 in the presence of a second molecule of amine, obtaining an activation enthalpy (40 kJ mol−1) which was in good agreement with the experimental value. The reaction product was described as the H-bond-stabilized adduct of dimethylcarbamic acid with dimethylamine. In contrast to aliphatic amines, aromatic amines hardly react with CO2 [21a] because of their poorer basicity. However, in the presence of suitable auxiliary bases (B) (such as amidines or penta-alkylguanidine superbases), carbamate salts (BH)O2CNRAr (R = H, alkyl) can be generated in solution, as supported by spectroscopic and reactivity data [29]. It has been shown that even tributylamine may be effective if a suitable alkali metal salt is also present in solution; in the latter case, the N-arylcarbamate has been isolated as an alkali salt (Equation 6.3) [3j, k].

6.3 Synthesis of Carbamate Esters

The reaction of 15N-labeled primary amines with 13CO2 in the presence of a strong base such as N-cyclohexyl-N′,N′,N″,N″-tetraethylguanidine (CyTEG) deserves attention. In CH3CN, in the presence of two equivalents of CyTEG, Et15NH2 underwent double carboxylation with formation of Et15N(13CO2)22−(+HCyTMG)2, as evidenced by means of 15N and 13C NMR spectroscopy. While with Ph15NH2 only Ph15NH13CO2− + HCyTMG was formed, with Cy15NH2 both species, Cy15NH13CO2− +HCyTMG and Cy15N(13CO2)22−(+HCyTMG)2, were observable, although the ratio of the two species was heavily dependent on the temperature at which the spectra were obtained [29b]. Nowadays, interest in the chemistry of amines (including ammonia) with CO2 remains very high, mainly because of its potential applications. Besides the traditional uses of amine chemistry in the chemical industry (e.g., Solvay process, synthesis of urea from NH3 and CO2) [30], new applications have begun to emerge, a few of which have synthetic relevance and are devoted to the synthesis of carbamates, isocyanates, and ureas (see Sections 6.3–6.5). However, other applications are more closely related with the reversible character of this chemistry. Alkylammonium carbamates are thermally unstable and release CO2 upon heating (Scheme 6.1), a fact which may prove useful under several circumstances. Both amines and polymer-bound amines have been investigated as reusable “CO2 scrubbers” for removing CO2 from industrial exhaust streams [31]. To this end, ionic liquids incorporating –NH2 groups in their structure have been developed, studied and proposed as attractive systems for gas puriﬁcation [32]. This method has been also extended to the use of multiple amine-containing dendrimers [33]. Dynamic CO2-amine chemistry has been successfully exploited for the reversible formation of organogels by exposing solutions of long-chain alkylamines to CO2 [34], or to build up supramolecular polymers [35]. Recently, thermally reversible carbamate chemistry has been employed also for the molecular imprinting of polymers [36] and for preparing switching-polarity solvents [37].

6.3 Synthesis of Carbamate Esters

Carbamate esters are very useful products. The carbamic group characterizes the molecular structure of several compounds which ﬁnd application in pharmacology (as drugs and prodrugs) [38] or as agrochemicals [39]. Organic carbamates also play a key role in synthetic chemistry as suitable intermediates for protecting amino-groups [40a], as linkers in combinatorial chemistry [40b], or as the precursors of ureas, isocyanates, and polymers [6–8, 40c]. The most common industrial syntheses of carbamate esters are based on either the alcoholysis of phosgene, followed by aminolysis of the intermediate chloroformate, or the reaction of an alcohol with an isocyanate, usually prepared from COCl2 [6, 7]. The development of phosgene-free routes to carbamates represents an important synthetic challenge, and several synthetic strategies are currently under investigation in this area.

125

126

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

E (b) CO2 evolution: side-products

R R'

CO2 incorporation: carbamate ester

O N

(b')

(a)

C O

(a')

carbamate moiety Scheme 6.3

Interaction of the carbamate group with an electrophilic organic substrate (E).

A few of these imply the carbonylation of nitroaromatic substrates [41], or the oxidative carbonylation of amines [42], or the reaction of amines with carbonic acid diesters [11] that are currently available, even on an industrial scale, through phosgene-free routes [43]. Another intriguing synthetic approach is based on the direct use of CO2 [44]. This route requires the transfer of carbamic group from carbamate sources (see Section 6.1), as metal-, alkylammonium, and p-block carbamates, to a suitable electrophilic organic substrate. The transfer reaction is not a banal step; rather, the carbamate anion, RR′NCO2−, is a remarkable case of an ambident nucleophile, as it shows two sites for interacting with an electrophile, E (Scheme 6.3), namely the oxygen ends, and the carbamic nitrogen atom. The formation of a carbamate ester implies that the electrophilic attack by E is to be directed towards one of the O-atoms (Scheme 6.3, (a)), and this results in the incorporation of originally ﬁxed CO2 into the RR′NCO2 group (Scheme 6.3, (a′)). Conversely, an electrophilic attack towards the nitrogen atom (Scheme 6.3, (b)) induces the elimination of CO2 from the carbamic moiety and the formation of undesired side products (Scheme 6.3, (b′)). To date, several organic substrates have been investigated; an overview of this research is presented, and its recent advances reviewed in the following sections. 6.3.1 Transfer of Carbamate Group to Alkyl Halides

Alkyl halides have been extensively investigated as electrophilic substrates in carbamation reactions of amines with CO2. Despite the great number of metal-carbamates isolated to date, the transfer of a carbamic group from a metal center to an alkylating agent such as an alkyl halide or dialkyl sulfate has been documented in only a few cases [1, 2]. In some initial studies [45], Saegusa and coworkers reported the synthesis of methylurethanes by the reaction of copper(I)-carbamates with methyl iodide. Here, the carbamate formation occurred only in the presence of a bulky σ-donor ligand,

6.3 Synthesis of Carbamate Esters

such as CNBut or P(Bun)3. This suggested that the reactivity of metal carbamates towards alkyl halides could be modulated by suitably changing the nature of the coordination sphere around the metal center. Accordingly, the reactivity of alkali carbamates (M = Li, Na, K) with alkyl halides can be addressed towards carbamation, rather than N-alkylation, if a suitable complexing agent (crown-ether, criptand) for the alkali metal cation was used [3j, 5a–c, 46]. In the absence of a complexing agent, the strong interactions between the alkali cation and carbamate anion depress the nucleophilicity of carbamic oxygens and alkyl halides react with the nitrogen atom, despite the delocalization of the nitrogen lone-pair on the carboxyl group. The complexing agent by spreading and shielding the positive charge of cation originates much more loose interactions between the cation and the O-ends of carbamate anion, the O-nucleophilicity of which is thus increased. Thus, anion–cation interactions seem to play a key role in this chemistry. Alkali carbamates can be obtained in a number of ways [1, 2, 3j]. Aresta and Quaranta have shown that alkali carbamates M(O2CNR2) (M = Li, Na, K) can be prepared, under mild conditions, from phosphocarbamates P(NR2)(O2CNR2)2 by reaction with an alkali halide MY in the presence of a suitable crown-ether, as complexing agent [5a–c]. If an alkyl halide R′X is present in the reaction mixture, then the transfer of a carbamic group from the phosphocarbamate to the alkyl halide can be successfully accomplished with a yield and selectivity that depend on the alkali salt MY used. The highest yields were obtained by using KF. The process, which is summarized in Scheme 6.4, remains the unique example of the utilization of p-block carbamates as a source of carbamate groups in the synthesis of carbamate esters. The transfer is mediated by intermediate alkali carbamate K(18-crown-6)O2CNR2, acting as a carrier of the carbamate group. The crown-ether can be recovered in practically quantitative yield; a complete recycling of P(NR2) F2 to obtain the starting tris-dialkylaminophosphine could not be accomplished in quantitative yields. The accomplishment of this step in 100% yield would be of great interest, as it would not only make the process fully cyclic but also provide potential applications. Yoshida et al. were the ﬁrst to report the synthesis of carbamate esters by the direct reaction of aliphatic amines, CO2 and alkyl halides [47]. The process involved the O-alkylation of intermediate alkylammonium carbamate salt, and required relatively, severe conditions (333–393 K; 4 MPa CO2), long reaction times (1–2 days) and an excess of amine (2.5 equiv.) with respect to the alkylating agent. The method was shown to be effective only with secondary aliphatic amines which, however, were converted into organic carbamates in low to moderate yield and with modest selectivity because of signiﬁcant side-formation of N-alkylation products. Aresta and Quaranta focused on the direct synthesis of carbamate esters by the reaction of aliphatic primary amines with alkyl halides in the presence of CO2. Under the conditions used (293–353 K; 0.1 MPa CO2; solvent = tetrahydrofuran (THF), MeOH, PhCH3/CH2Cl2 mixtures), the formation of carbamate ester was

127

128

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas PX3

R2NH

X=Cl, Br,I

R2NH2X BH+F-

P(NR2)3 CO2

B, R2NH P(NR2)F2

P(NR2)(O2CNR2)2

KF

[ K.(18-crown-6)] O2CNR2

18-crown-6

R'X

R2NC(O)OR' YIELD > 90%

KX [K .(18-crown-6)]X

R = Me, Et R' = CH2Ph, CH2CH=CH2, n-C10H21

Overall carbamation reaction: P(NR2)(O2CNR2)2 + 2 R'X + 2 KF + 2 (18-crown-6) . (18-crown-6)]X 2 R2NC(O)OR' + P(NR2)F2 + 2 [K

Scheme 6.4

Synthesis of organic carbamates from phosphocarbamates.

not observed at all, or it may have occurred only to a very minor extent. The reaction afforded mainly N-alkylation products, which may form by the direct alkylation of free amine present at equilibrium, or they may be the result of an electrophilic attack by alkyl halide at carbamic nitrogen atom of carbamate anion

6.3 Synthesis of Carbamate Esters

or other carbamic species (RNHCO2H, [RNHC(O)OH]2; see also Scheme 6.1) present in the reaction mixture [27]. The synthesis of carbamate esters RNHC(O) OR′ from primary amines, CO2 and R′X implies O-alkylation of the intermediate ionic carbamate (RNH3)O2CNHR, and requires that the reactivity of (RNH3) O2CNHR towards R′X must be modiﬁed. This was achieved by working in the presence of a suitable macrocyclic polyether (18-crown-6-ether) [27], where Oalkylation could occur in competition with N-alkylation and organic carbamates were obtained in satisfactory yields at ambient temperature. On completion of the reaction, the crown-ether could be recovered in quantitative yield, and recycled. The change of reactivity has been ascribed to the formation of a “host–guest” adduct (Scheme 6.5) between the crown-ether and the alkylammonium cation RNH3+. The crown-ether can both increase the solubility of carbamate salt (which usually is poorly soluble in most organic solvents) and also change the reactivity of carbamate anion as an ambident nucleophile. Ionic association phenomena, which are due to hydrogen bonding between the alkylammonium cation and the carbamate anion, can strongly depress the O-nucleophilicity of these salts. The

N-Alkylation

RNH2 +

CO2

R'X

18-crown-6

O-Alkylation: carbamate esters

2 RNH2 + CO2 + 18-crown-6 + R'X

RNHCO2R' + [RNH3(18-crown-6)]X

(a)

Mechanism: [RNH3(18-crown-6)]O2CNHR

2 RNH2 + CO2 + 18-crown-6

[RNH3. (18-crown-6)]O2CNHR + R'X

RNHCO2R' + [RNH3(18-crown-6)]X

O

R

(b) (c)

O

H

RNHC(O)O

-

O

N

+

H

O

H

O

O

Scheme 6.5 Direct synthesis of carbamate esters from primary amines, CO2, and alkyl halides in the presence of 18-crown-6-ether.

129

130

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

complexation of RNH3+ ions by crown-ether molecules weakens the anion–cation interactions and enhances the O-nucleophilicity of the carbamate anion, as already discussed for Group 1 metal carbamates. Carbamate esters have been prepared, under mild conditions, using suitable sterically hindered organic bases B, such as amidines [29, 48], penta-alkylguanidines [29b–d], and phosphazenes [49]. The base B must be strong enough to drive formation of the carbamate salt (BH)O2CNRR′, and also able to enhance the O-nucleophilicity of carbamate ion by generating a highly polarizable counterion BH+, which can favor the formation of naked RR′NCO2− anions [29, 48, 49]. Simple tertiary amines are poorly effective. Although a remarkable increase in carbamate yield can be achieved by using DBU (1,8-diazabicyclo[5.4.0]undec-7-ene) [29, 48]. Higher carbamate selectivities were reported with penta-alkylguanidine superbases, such as CyTMG [29b–d]. Penta-alkylguanidines are sterically hindered bases that are stronger than amidines and able to produce, upon protonation, cations with a greater delocalization of the positive charge. These properties favor a higher concentration of carbamate anions in solution, as well as a more effective ionic separation which makes the O-end of carbamate ion more available to electrophilic attack. Inorganic bases have been also employed in this system. When Butcher ﬁrst used alkali carbonates [50], it was reported that, in DMF and at ambient temperature, the carbamation of primary and secondary aliphatic amines (or also arylamines) with alkyl halides under a CO2 atmosphere (0.1 MPa) was effectively promoted by Cs2CO3 [50, 51]. The Cs+ cations in the solvent used (DMF) did not form ion pairs with counterions, and favored the formation of naked carbamate anions that were more reactive at the O-ends with alkyl halides. Jung further found that the addition of tetrabutylammonium iodide (TBAI) to the system RR′NH/ CO2/RX/Cs2CO3/DMF promoted the carbamation process with a higher yield and selectivity with respect to N-alkylation [51]. The process has been successfully extended to the synthesis of carbamate functionalities on solid phases. In this case, resin-bound carbamates are readily released from the resin by treatment with LiAlH4 in THF, yielding the respective N-methyl secondary amines [51]. Recently, organic carbamates have been synthesized by the reaction of amines and alkyl halides with scCO2 in the presence of potassium carbonate and an onium salt (Bu4N)Br [52]. In order to ascertain the carbonyl-active species (carbonate or CO2?), the reaction was carried out also with potassium phosphate instead of the carbonate salt; the similar results obtained conﬁrmed that scCO2 acted in the process not only as a solvent but also as a carboxylating agent. Other base systems have been also studied, including tetraethylammonium superoxide [53], basic resins [54], Triton-B (benzyltrimethylammonium hydroxide) [55], and K2CO3 in the presence of catalytic amounts of (Bu4N)I [56]. In a few cases, the use of alkyltosylates, as alkylating agents in place of alkyl halides, has been also investigated [54b, 55b, 56]. The base can also be generated electrochemically by the cathodic reduction of: (i) O2 to superoxide O2− in the presence of CO2 [57a]; (ii) a suitable probase such as pyrrolidin-2-one [57b]; and (iii) the supporting electrolyte-solvent system, as in

6.3 Synthesis of Carbamate Esters

the case of CH3CN-TEAP (tetraethylammonium perchlorate) [57c]. The electrochemical activation of CO2 to radical anion CO2−., in either conventional solvents [58a, b, d] or in ionic liquids [58c], also promotes the formation of carbamate esters from aromatic or aliphatic amines and alkyl halides under mild conditions. With BMImBF4 (BMIm = 1-butyl-3-methyl imidazolium) as solvent, the carboxylation of amines to carbamate anion has been related to the formation of basic CO32− anion, obtainable via CO2−./CO2 coupling reactions [58c]. The radical anion CO2−. was also found to promote the carboxylation of N-acyl or N-alkoxycarbonyl alkylamines to give cyclic carbamates [57a, 59]. The use of heterogeneous catalysts in the synthesis of urethanes from aliphatic and aromatic amines, CO2 and alkyl halides has been explored only recently. Titanosilicate molecular sieves [60a], metal phthalocyanine complexes encapsulated in zeolite-Y [60a], beta-zeolites and mesoporous silica (MCM-41) containing ammonium cations as the templates [60b, c], and adenine-modiﬁed Ti-SBA-15 [60d, e] each function as effective catalysts, even without any additional base. 6.3.2 Transfer of Carbamate Group to Acylating Agents

The transfer of carbamate groups from metal carbamates to acylating agents, such as acyl halides and chloroformates, has been extensively studied by Calderazzo and coworkers. At ambient temperature, the reaction of metal N,N-dialkylcarbamates with acyl halides, R′C(O)Cl, or acetic anhydride, yields a carbamic-carboxylic mixed anhydride (Equation 6.4) which, on occasion, may undergo decarboxylation and convert into the corresponding amide R2NC(O)R′ [2]. M (O2CNR 2 )n + nR ′C (O) X → MX n + nR 2NC (O) OC (O) R ′

(6.4)

X = Cl, M = Na, Cu(II), Mn(II), Co(II), Ti(III), V(III), Fe(III) X = OC(O)CH3, M = Ni(II), “(HNEt2)2Pd(II)” The reaction of metal N-alkylcarbamates M(O2CNHR)n (M = Na, Mn(II), Co(II); R = Ph, Pr, Cy) with R′C(O)Cl (R′ = Me, Ph) takes place, at ambient temperature, in a more complex way with the formation of isocyanates (RNCO), carboxylic anhydrides (R′C(O)OC(O)CR′), amides (RNHC(O)R′) and CO2. Amide formation and the evolution of CO2 can be due to: (i) the decomposition of mixed anhydride RNHC(O)OC(O)R′ obtained by addition of the acyl chloride to the oxygen atom of the carbamate group; or (ii) the direct reaction of acyl chloride at the carbamic nitrogen atom of M(O2CNHR)n. The mixed anhydride RNHC(O) OC(O)R′ might also decompose via another route so as to afford isocyanate and carboxylic acid. However, a different pathway (Scheme 6.6) has been also envisaged for the formation of RNCO and R′C(O)OC(O)CR′, which excludes any intermediacy of the mixed anhydride [61a]. Two acetic acid molecules, bound to the same metal or to different metal centers, would then be dehydrated and acetic

131

132

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

C

N H

O Cl

R

R

O M

Cl

O=C=N

O

O

C

+

M

Me

C OH

Me

Scheme 6.6 isocyanate.

Reaction of metal carbamates with acylating agents: a plausible route to

acid anhydride released into solution, while the water would remain bound to the metal. Mn6(O2CNEt2)12 reacted with phosgene and chlorofomates according to Equation 6.5. The electrophilic attack takes place at the carbamic oxygen, with formation of the primary product Et2NC(O)OC(O)X (X = Cl, OR). A mechanistic study of the reaction, using Mn6(O213CNEt2)12, has shown that the formation of Et2NC(O)X from Et2N13C(O)OC(O)X involves cleavage of the N–13C bond [61b]. 1 6 [Mn 6(O2CNEt 2 )12 ] + 2ClC (O) X → MnCl 2 + 2CO2 + 2Et 2NC (O) X

(6.5)

where X = Cl, OEt. Aresta and Quaranta studied the reactivity of alkylammonium N-alkylcarbamates (RNH3)O2CNHR towards a different acylating substrate, such as dimethyl carbonate (DMC) [62a, b]. Carbamate salts (RNH3)O2CNHR (R = benzyl, allyl, cyclohexyl), prepared in situ from aliphatic primary amines and CO2, reacted with DMC to afford N-alkyl methylcarbamates (Equation 6.6). The reaction requires mild conditions (343–363 K; 0.1 MPa CO2 pressure) and can be carried out in DMC used as solvent and reagent. At 363 K, carbamate esters were obtained in satisfactory yield (45–92%) with high selectivity, as side products such as ureas, N,N-dialkylcarbamate esters, and alkylated amines were formed in very small amounts.

(RNH3 ) O2CNHR + (MeO)2C = O → RNHC (O) OMe + MeOH + CO2 + RNH2

(6.6)

The reaction in Equation 6.6 involves, as the key step, the O-methoxycarbonylation of carbamate anion to give a carbamic-carbonic mixed anhydride, RNHC(O)OC(O) OMe (Equation 6.7) [62d], which, under the working conditions, undergoes a very rapid decarboxylation and is converted into the carbamate product (Equation 6.8).

(RNH3 ) O2CNHR + (MeO)2C = O → RNHC (O ) OC (O) OMe + MeOH + RNH2 RNHC (O) OC (O) Me → RNHC (O) OMe + CO2

(6.7) (6.8)

Subsequent labeling experiments with 13CO2 showed that the 13CO2 molecule originally ﬁxed by the starting amine was not incorporated in the ﬁnal product, but rather was released when the organic carbonate was formed (Scheme 6.7). Therefore, CO2 can act as a catalyst for amine methoxycarbonylation with DMC (Equation 6.9 in Scheme 6.7) [11]. The use of CO2 as catalyst for Equation 6.9

6.3 Synthesis of Carbamate Esters *

RNHC(O)OMe

CO2

*

RNH2

*

RNH2 CO2 .

RNHC(O)OC(O)OMe

MeOH RNH2

RNH2 DMC

Overall process: OMe RNH2 + O

[RNH3][O2*CNHR]

CO2

RNHC(O)OMe + MeOH (6.9)

OMe

Scheme 6.7 Synthesis of carbamates from aliphatic primary amines and dimethyl carbonate (DMC) in the presence of CO2. The catalytic role of CO2.

(Scheme 6.7) is particularly attractive, as this species is inexpensive, nontoxic, and does not present regeneration and recycling problems. Kinetic studies have shown that the overall Equation 6.9 follows a second-order kinetic law for a wide number of primary aliphatic amines RNH2 (R = PhCH2, Cy, n-propyl, allyl, sec-butyl, iso-butyl, n-butyl) [62c]. The rate constants and activation energies clearly showed the existence of an isokinetic effect, and have been used as input data for a theoretical study of the reaction mechanism according to the “Selective Energy Transfer (SET) model” [62c]. The application of this model to Equation 6.9 supported that the rate-determining step of the process was the formation of a carbonic-carbamic mixed anhydride by reaction of the ion pair RNH3+ −O2CNHR with DMC (Equation 6.7). As the ion pair is approaching the organic carbonate, a hydrogen bond between the alkylammonium cation and the nearest oxygen of DMC is formed. This H-bond opens up a reactive pathway for the splitting off of a molecule of methanol and the simultaneous release of a molecule of the amine (Scheme 6.8). This synthetic methodology was extended to the carbamation of some industrially relevant aminofunctional silanes, such as H2N(CH2)3Si(OMe)3, H2N(CH2)3Si(OEt)3, H2N(CH2)2NH(CH2)3Si(OMe)3 [23a], which were selectively methoxycarbonylated and converted (343–348 K, PCO2 = 0.1 MPa), in high yield, into the corresponding carbamates MeO(O)CNH(CH2)3Si(OMe)3, MeO(O)CNH(CH2)3 Si(OMe)x(OEt)3–x and MeO(O)CNH(CH2)2NH(CH2)3Si(OMe)3, respectively. The carbamates MeO(O)CNH(CH2)3Si(OMe)3, MeO(O)CNH(CH2)3Si(OMe)x(OEt)3–x and MeO(O)CNH(CH2)2NH(CH2)3Si(OMe)3 are useful silane coupling agents, and

133

134

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

H

R N

O - O

CH3

H

O

+ O

H3C O Scheme 6.8 (DMC).

H

H

N R

Interaction of alkylammonium carbamate ion pair with dimethyl carbonate

serve as a ready source of isocyanates that are largely used in chemical industry as modulators of the physico-mechanical properties of polymeric materials. 6.3.3 Transfer of Carbamate Group to Alcohols

With respect to the synthesis from amines, CO2 and alkyl halides, the synthesis of carbamates from amines, CO2 and alcohols (Equation 6.10) is not only a phosgene-free, but also a halogen-free process. Moreover, water forms as the only reaction coproduct. Whilst these features make the route very attractive from the point of view of environmental sustainability, unfortunately the reaction suffers from both thermodynamic and kinetics limitations. Kinetic impediments make necessary the use of a suitable catalyst which, moreover, must be water-tolerant in order to avoid deactivation by cogenerated H2O. Several strategies have been explored to overcome these restraints, based mainly on the use of alcohols in a dehydrated form (for instance, as ortho esters or ortho carbonates) [63], or on the use of dehydrating agents [64, 65]. Cat . ⎯⎯⎯ ⎯⎯ → RNHCO2R ′ + H2O RNH2 + CO2 + R ′OH ← ⎯

(6.10)

Sakakura has reported the synthesis of urethanes from sterically hindered primary aliphatic amines (t-BuNH2, CyNH2), CO2 and alcohols (MeOH, EtOH), using ketals (2,2-dimethoxy- or 2,2-diethoxypropane) as dehydrating agents [64]. The process was carried out in dense CO2 (30 MPa) under drastic conditions (473 K, 24 h) and catalyzed by tin derivatives (OSnBu2, Me2SnCl2, Bu2Sn(OMe)2) [64a] or also by less-toxic homogeneous Ni(II) complexes with bipyridine or phenanthroline or carbene ligands [64b, c]. De Vos and coworkers have shown that base catalysts can also promote the effective conversion of a large variety of amines and alcohols into carbamates by reaction with CO2 [65]. The most active among the catalysts tested were Cs2CO3 and Rb2CO3, which were able to convert both linear and branched aliphatic primary amines into their corresponding carbamates under relatively less severe conditions (473 K, 2.5 MPa), even in the absence of dehydrating agents. Only with a sterically

6.3 Synthesis of Carbamate Esters

hindered amine, such as tert-butylamine, was the use of a dehydrating agent (ketals, molecular sieves) proved to be necessary. Mechanistic studies have revealed two main reaction pathways for carbamate formation: (i) a direct pathway, with an isocyanate as reaction intermediate; and (ii) an indirect route, involving urea alcoholysis. In general, the reaction in Equation 6.10 requires quite severe conditions. Milder reaction conditions can be employed by using methanesulfonic anhydride, as a condensating agent, in the presence of bases [66]. A Mitsunobu-based protocol, based on Ph3P/diethylazodicarboxylate (DEADC) as an auxiliary reagent system, has been shown to be very effective for the carbamation under mild conditions (363–373 K, atmospheric CO2 pressure) of aliphatic and aromatic amines with a variety of primary, secondary, and tertiary alcohols [67]. However, the latter approaches, though useful from a merely synthetic point of view, are less attractive than the direct approach (Equation 6.10), mainly because high-energy reactants are required. This increases costs signiﬁcantly and markedly decreases the atom economy of the overall process, which results in the unavoidable coformation of stoichiometric amounts of wasted coproducts. One transformation that is closely related to the synthesis of carbamate esters from alcohols, amines and CO2, is the formation of cyclic urethanes, such as oxazolidin-2-ones, by reaction of CO2 with 1,2-amino alcohols or their formally dehydrated derivatives, aziridines. The incorporation of CO2 into the aziridine ring to give oxazolidin-2-ones has long been known. In fact, an early report described the direct synthesis of oxazolidin-2-ones (5–80% yield) from aziridine or 2-methylaziridine and CO2, in the presence of iodine as catalyst, at 333–353 K, under a pressure of CO2 [68a]. More recently, however, Kawanami and Ikushima revisited the catalytic process using scCO2, and found that the nature of the substrate greatly affected the regioselectivity of the process [68b]. When starting from 2-methylaziridine, only 4-methyloxazolidin-2-one was obtained, whereas 2-phenylaziridine was converted into 5-phenyloxazolidin-2-one (Scheme 6.9).

sc-CO2 (11.6 MPa), I2, MeCN, 313 K, 6 h

NH

N H

O

sc-CO2 (10.2 MPa), I2, MeCN, 313 K, 6 h

Ph N H

O 72% NH

Ph

O

O

76%

Scheme 6.9 Synthesis of oxazolidin-2-ones from 2-substituted aziridines and scCO2, in the presence of iodine. The regioselectivity of the CO2 insertion reaction.

135

136

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

The use of a hybrid reaction system which consisted of scCO2 and a room temperature ionic liquid (IL), based on a tetraalkyammonium salt, led to a dramatic acceleration of the conversion of 2-methylaziridine into 4-methyl oxazolidin-2-one in the presence of iodine, with yields of up to 98% being achieved in very short times (5 min; 313 K; CO2 pressure 10 MPa) [68c]. A very efﬁcient catalytic system has been developed by He and coworkers using, under solvent-free conditions, quaternary ammonium bromide-functionalized polyethylene glycol (0.25 mol%; 373 K, 8 MPa, 5–20 min) [68d]. The catalyst worked well for a variety of 1-alkyl-2-aryl-substituted aziridines, and could be easily recovered by centrifugation and reused, without any signiﬁcant loss of catalytic activity and selectivity. The ﬁxation of CO2 into a three-membered ring has also been promoted, under relatively severe conditions, by other catalytic systems such as tetraphenylantimony halides (333 K, 5 MPa) [68e] or the p-methoxyphenol/DMAP system (DMAP = 4,6-(dimethylamino)pyridine; 393 K, 3.6 MPa, 48 h) [68f] or, under much milder conditions, by alkali [68g–k] or tetralkylammonium halides [68j], or by (Salen)Cr(III)(DMAP) [68l]. It is worth noting that, with 2-alkyl or 2-aryl substituted aziridines, alkali or tetralkylammonium halides catalyze the formation of the 4-substituted regioisomer as the main or unique product, whereas the chromium(III) catalyst promoted the preferential conversion to the 5-substituted regioisomer with high selectivity and yield. Dunach has reported the insertion of the heterocumulene into N-Boc (Boc = t-butoxycarbonyl) -protected 2-substituted aziridines under mild conditions (ambient temperature, atmospheric CO2 pressure), using electrochemical methods and dibromo(1,4,8,11-tetraazocyclotetradecane)-nickel(II), NiBr2(cyclam), as catalyst (10 mol%) [68m]. However, the process was poorly selective as a mixture of 4- and 5-regioisomers was obtained in all cases investigated. Oxazolidin-2-ones can also be obtained by the reaction of CO2 with 1,2-amino alcohols. Under severe conditions of temperature and CO2 pressure, a few amino alcohols are converted into the corresponding cyclic urethanes, even in the absence of any catalyst [69]. The reaction can proceed in common solvents, including water, and also under solventless conditions. However, the yields and selectivities depend heavily on the structure of the amino alcohol used, and the experimental conditions. The cyclourethanization of 1,2-amino alcohols with CO2, in an aprotic medium (e.g., benzene), was effectively catalyzed by organoantimony compounds, such as triphenylstibine oxide [70a]. Molecular sieves 3A can assist Ph3SbO catalysis, which is sensitive to hydrolysis by water formed in situ. The catalytic system consisting of triphenylstibine oxide (10 mol%) and molecular sieves 3A synergistically promoted the formation of oxazolidin-2-ones in good yield and selectivities. Thus, a variety of 3-alkyl-2-oxazolidinones were prepared in 80–94% yields at 403–433 K (5 MPa CO2 pressure; 6–24 h) and trans-hexahydro-3-methylbenzoxazolidinone was obtained stereoselectively from trans-2-methylaminocyclohexanol. Unfortunately, under these working conditions, the catalytic system failed to promote the reaction of 2-aminoethanols without any N-substituent.

6.3 Synthesis of Carbamate Esters

The dehydrative condensation of 1,2-amino alcohols with CO2 was found to be catalyzed also by n-Bu2SnO (10 mol%) in N-methylpyrrolidone (NMP) as solvent [70b]. After 16 h at 453 K (CO2 = 5 MPa), 2-oxazolidinones were obtained in 53– 94% yields. In the solvent used (NMP), in the absence of any catalyst, even N-unsubstituted 1,2-amino alcohols gave the corresponding cyclic carbamate in fair yield, although the addition of n-Bu2SnO further improved yields by more than 50%. Recently, several ILs, in the presence of alkali metal compounds as promoters and ethanol as solvent, have been investigated as catalysts [70c]. The best catalytic activity was exhibited by the system BMImBr/K2CO3. However, under the working conditions (423 K; 6 h; CO2 = 6–10 MPa; IL = 19 mol%), the carbamate yields and selectivities were modest because of the side-formation of cyclic ureas, and oligomeric byproducts. The use of condensating reagents, such as tetraphenylpyrophosphate [71], phosphinic acid anhydride[71], carbodiimides [72], tosyl chloride [73], and a variety of deoxygenating agents has been explored [74–76]. These systems have been shown to promote the cyclourethanization reaction in good yield under relatively milder conditions. Kodaka and coworkers prepared oxazolidin-2-ones from amino alcohols derivatives and CO2 (atmospheric pressure), at ambient temperature, in the presence of triethylamine as base and using DEADC/PR3 as Mitsunobu’s reactants in CH3CN [75a]. It has been proposed that the synthesis of cyclic urethane may proceed according to the mechanism shown in Scheme 6.10. Using C18O2 and (S)-2-benzyl2-aminoethanol (Scheme 6.10; R3 = H; R1 = benzyl; R2 = H), it has been found that the ﬁnal step depends on the phosphine used. With PPh3, unlabeled phosphine oxide Ph3P=16O was mainly obtained, thus indicating that the hydroxyl group of the amino alcohol was activated. On the other hand, with tributylphosphine, labeled phosphine oxide, Bu3P=18O, was produced, which was consistent with activation of the carbamic moiety of (a), followed by the nucleophilic attack by hydroxyl group to the carbamic carbon atom. Dinsmore and Mercer further investigated this reaction using DBU as a base and n-Bu3P/DBAD (di-tert-butyl azodicarboxylate) as Mitsunobu’s reactants, and found an unexpected steroselectivity in the Mitsunobu transformation [75b]. In fact, the stereochemical course of the Mitsunobu reaction (Scheme 6.11) depended on whether the carbamic acid intermediate was N-substituted with hydrogen (retention) or with carbon (inversion). The ring closure of amino alcohols with CO2 to yield cyclic carbamates, was achieved under mild conditions (atmospheric pressure of CO2, room temperature), in acetonitrile as solvent and in the presence of triethylamine as the base, using more easily available reactants, such as P(III)-derivatives (Ph3P, (PhO)3P, n-Bu3P, (MeO)3P) and haloalkanes (CCl4, CCl3CCl3) [76]. According to the proposed mechanism, the active species is a phosphonium adduct of the used P(III)-compound with the haloalkane, which activates the intermediate carbamate formed from amino alcohol and CO2 at the carbamic moiety to produce a transient species which cyclizes to the ﬁnal product (Scheme 6.12).

137

138

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas 3

R NH

1

R

3

R * NCO 2

1

R * + NEt + CO 2 3

2

R

+

Et3NH R2

OH

OH (a)

EtO2CN=NCO2Et + PR3

EtO2CN-NCO2Et + PR3 (b)

3

R1

R * + NCO2PR3

R = n-Bu R2 (a) + (b)

1

R2

R

O

N R3

OH

* + OPR3

*O

3

1

R = Ph

R

R * NCO

2

R1

2

R

3

*O 2

R

+ O-PR3

N R

+ OPR3

*O

Scheme 6.10 Proposed mechanism for the synthesis of oxazolidin-2-ones from amino alcohol derivatives and CO2 using DEADC/PR3 as Mitsunobu’s reactants and triethylamine as the base (solvent: CH3CN).

6.3.4 Transfer of Carbamate Group to Epoxides

The transfer of a carbamate moiety to epoxides is a suitable entry into carbamic esters of 1,2-diols. Yoshida and Inoue reported that the reaction of Ti(NMe2)4 with CO2 and 1,2-epoxycyclohexane, followed by hydrolysis of the reaction mixture, afforded trans-2-hydroxycyclohexyl N,N-dimethylcarbamate [77a]. This reaction was the ﬁrst example of transfer of a carbamate group from a metal carbamate to an epoxide (Scheme 6.13). Other metal-amides, such as TiCp(NMe2)3, W2(NMe2)6 and W(NMe2)6 [77b], or EtZn(NPh2) and Et2Al(NPh2) [77c], have also been used in the synthesis of trans2-hydroxycyclohexyl N,N-disubstituted carbamates through the same route. Monocarbamates of 1,2-diols can be obtained also by the direct reaction of epoxide with primary or secondary aliphatic amines in the presence of CO2

6.3 Synthesis of Carbamate Esters

O HO

NH2

O

CO2, DBU, MeCN Ph

Ph

NH

n-Bu3P, DBAD Ph

Ph 96 %

N substituted with H (retention) O HO

HN-CH2Ph CO2, DBU, MeCN

NCH2Ph

O

n-Bu3P, DBAD Ph

Ph Ph

Ph 84 %

N substituted with C (inversion) Scheme 6.11 Stereochemical course (retention versus inversion) for the synthesis of oxazolidin-2-ones from amino alcohol derivatives and CO2 using DBU as the base and n-Bu3P/ DBAD as Mitsunobu’s reactants, in CH3CN.

1

2

R

R NH

H 3 R R3P + CCl4

R3P

+

CCl3 Cl

H

1

2

OH

CO2 Et3N

R H 3 R

H

R N

O C O

O R P 3

+ CHCl3 + Et3NHCl

1

2

R R3P=O +

R N

H 3 R H

C O

Scheme 6.12 Proposed mechanism for the synthesis of cyclic carbamates from amino alcohols and CO2, using P(III) reagents and haloalkanes.

O

139

140

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

Ti(NMe2)4 + CO2 +

OH

H2O

O

O O C NMe2

O

O Ti(NMe2)4 + CO2

Ti(OCONMe2)4

Ti

O

O C NMe2

4

Scheme 6.13

Transfer of carbamate group from a metal carbamate to epoxides.

R' CHCH2 O R' CH3CO2 Al O2CNR2

+

R2N H2

CH3CO2 Al OCHCH2O2CNR2

R2N+H2

C

A R' HOCHCH2O2CNR2

CO2 + HNR2 (HO2CNR2)

B

Scheme 6.14 Proposed mechanism for the formation of hydroxycarbamates from secondary aliphatic amines, CO2 and epoxides catalyzed by Al(TPP)(O2CCH3) [(5,10,15,20tetraphenylporphyrinato)aluminum(III) acetate].

[78, 79]. The reaction with unsymmetrical epoxides as propylene or styrene oxides produced a mixture of carbamate isomers, RR′NCO2CH(R″)CH2OH and RR′NCO2CH2CH(R″)OH (R″ = Me, Ph). The synthesis of hydroxycarbamates from secondary aliphatic amines, CO2 and epoxides has been found to be catalyzed by (5,10,15,20-tetraphenylporphinato) aluminum(III) acetate, Al(TPP)(O2CCH3) [80]. Scheme 6.14 illustrates the mechanism proposed for the catalytic process, which can be carried out under not severe conditions (293–343 K; 0.1–5 MPa CO2 pressure). The key step here is the insertion of epoxide into the Al–O bond of the Al-carbamate A (Scheme 6.14), which preliminarily forms by the reaction of Al(TPP)(O2CCH3) with the amine and CO2. Protolytic cleavage of the Al-alkoxide bond in the insertion product, C, by dialkylcarbamic acid regenerates the catalytically active carbamato-species A and

6.3 Synthesis of Carbamate Esters OH

OH ClCH2CH ClCH2CH

CH2

CH2

RHN

+ CO2 + NH2R

O N

C

O

O

R

O

O

Scheme 6.15 Synthesis of cyclic carbamates from chloromethyl oxirane, primary aliphatic amines, and CO2.

CH3 Ph

C

C

O

X

R'

O CH3O

Ph

C

CH3 C

OCH3

R'

R

O RNHCO NH3R

Ph

N

O C

C

O C

CH3

OH

R'

Scheme 6.16 Proposed mechanism for the reaction of CO2 with α-haloacylophenones and aliphatic primary amines, in methanol.

liberates the carbamate ester B. Remarkably, the formation of the other isomer of B, R2NCO2CH(R′)CH2OH, was not observed. With oxiranes suitably functionalized, cyclic carbamates can be obtained. Accordingly, when chloromethyl oxirane was used as the epoxydic substrate, its reaction with primary aliphatic amines and CO2 (333 K, 1 MPa, 10–15 h) gave 3-alkyl-5-hydroxyoxazin-2-ones, albeit in a low yield of 2.5–10% (Scheme 6.15) [79]. Both, ﬁve- and six-membered cyclic carbamates have been synthesized by the reaction of 2-(1-haloalkyl)-oxiranes with CO2 and primary aliphatic amines [81]. Notably, 5-substituted 2-oxazolidinones have been prepared in good yield (51–94%) by reacting 2-aminomethyloxiranes with CO2 (0.1 MPa), in MeOH, at room temperature [82]. The synthons of oxiranes have also been used in this respect. For example, the reaction of CO2 with α-bromoacylophenones in the presence of aliphatic primary amines, in methanol, afforded 3-alkyl-4-hydroxyoxazolidin-2-one derivatives under mild conditions [83a]. However, neither α-bromoacetophenone nor α-chloroacetophenone afforded any carbamate product, and no urethanes were obtained with aromatic or aliphatic secondary amines. The proposed mechanism involved, as the ﬁrst step, the formation of a 3-alkyl-2-methoxy-2-phenyloxirane intermediate, which reacted with alkylammonium carbamate to give the oxazolidone product (Scheme 6.16). This synthetic protocol was successfully applied to the synthesis of bis(oxazolidin-2-one) derivatives by reactions of 2-methoxy-3,3dimethyl-2-phenyloxirane or α-bromoisobutyrophenone with CO2 and aliphatic α,ω-diamines [83b]. The use of oxetanes has been also explored. Both primary and secondary aliphatic amines react with oxetanes, under CO2 (4 MPa), at 373–393 K, in the absence of any catalyst, to afford monocarbamates of 1,3-propanediols, together with amino alcohols as side products [84].

141

142

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

6.3.5 Transfer of Carbamate Group to C–C Double Bonds

The transfer of a carbamate moiety to oleﬁns has been documented only in very few cases, using activated unsaturated substrates. For example, Yoshida and Inoue reported the selective (100%) formation of 1-ethoxyethyl N,N-dialkylcarbamate esters by the reaction of CO2 (5 MPa) with ethyl vinyl ether and secondary amines R2NH (R = Me, Et) in the absence of any catalyst (Equation 6.11) [85]. R 2NH + CO2 + CH2 = CHOEt → R 2NCO2CH (CH3 ) OEt

(6.11)

In this reaction, carbamate esters were obtained after long reaction times (70 h), in 0.06–11% yield, depending on the oleﬁn/amine molar ratio (1–10 mol/mol) and temperature (308–353 K). Since 1-ethoxyethyl carbamate was formed as the sole regio-isomer product, but there was no formation of the 2-ethoxyethyl ester, the reaction was believed to proceed by an electrophilic addition to the vinyl ether of carbamic acid, formed by reversible reaction of CO2 with amine (see Equation 6.1). According to another approach, the oleﬁnic substrate was activated through coordination to a suitable metal center [86]. In fact, the addition of preformed carbamate anions RR′NCO2− (which had been generated from various primary or secondary amines and CO2) to (norbornadiene)PdCl2 (at 194 K), followed by the addition of DIPHOS (1,2-bis(diphenylphosphino)ethane), and subsequent reductive cleavage with NaBH4, gave nortricyclo carbamate esters (A, Scheme 6.17) in good yields (75–100%). Moreover, when (dicyclopentadiene)PdCl2 or (1,5-cyclooctadiene)PdCl2 were used as the oleﬁn source, the corresponding carbamates (B and C, respectively in Scheme 6.17) were obtained in good yield after reductive cleavage with NaBH4 or dihydrogen. It is worth noting that, with norbornadiene as the coordinating oleﬁn, the addition of a pre-made carbamate anion to a slurry of the palladium-complex at 194 K, followed by warming to ice temperature and addition of DIPHOS, afforded (DIPHOS)Pd(Cl)(nortricyclo urethane).

Ph

RR'NCO2

Ph P

Pd Cl

P Ph

Ph

(DIPHOS)Pd(Cl)(nortricyclo urethane)

Quenching of this species with anhydrous HCl also gave the nortricyclo carbamate, together with (DIPHOS)PdCl2. This result indicated that the Pd–C bond

6.3 Synthesis of Carbamate Esters

1) [R3NH][O2CNRR'] 2) DIPHOS 3) NaBH4

Pd Cl

Cl

1) [R3NH][O2CNRR']

A

RR'NCO2

2) H2 or NaBH4

Pd Cl

RR'NCO2

B

Cl

n n

n

1) [ BuNH3][O2CNHBu ] Pd Cl

Cl

BuNHCO2

2) NaBH4 C

Scheme 6.17 Metal-assisted carbamation of oleﬁns.

in (DIPHOS)Pd(Cl)(nortricyclo urethane) was too strong for protonolysis by the ammonium ion, although its cleavage could be more easily accomplished by a stronger acid, such as HCl. Taken together, these ﬁndings suggest that, with the correct choice of metal–ligand environment, the carbamation of oleﬁn might be carried out catalytically, as illustrated in Scheme 6.18. Pd(0)/phosphine complexes, or their precursors, in the presence of a suitable co-base, have also been shown to promote, in good yields (66–100%), the formation of allylic carbamates from various primary and secondary aliphatic amines, pressurized CO2 and allylic chlorides, in THF, at ambient temperature [87a]. The choice of the added co-base (Base), used for generating the carbamate salt RR′NCO2−(BaseH)+, was found to be critical for high yields of O-allylic urethanes. The use of a guanidine (CyTMG) or amidine (DBU) base was optimal for this system (see also Section 6.3.1). It is assumed that this chemistry passes catalytically through a mechanism similar to that illustrated in Scheme 6.19. This involves nucleophilic attack by carbamate anion on a (π-allyl)palladium species, formed by the oxidative addition of the allylic chloride to a palladium(0) intermediate. In a more recent report, (E)-4-(benzylamino)-2-butenyl methyl carbonates were converted into substituted vinyloxazolidinones through a CO2-recycling reaction catalyzed by Pd(PPh3)4 in the presence of DBU [87b]. The process was believed to involve the Pd-promoted decarboxylation of carbonate moiety of substrate with the

143

144

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

RR'NCO2

+ RR'NH

RR'NH2+ M-Ln

M-Ln RR'NCO2

M-Ln

RR'NH2+

-

O2CNRR'

2 RR'NH + CO2 Scheme 6.18

Metal-catalyzed carbamation of oleﬁns: a possible catalytic pathway.

X Nu

-

PdLn

Nu (nucleophile) -

X

X PdLn

Nu

Scheme 6.19 Palladium-catalyzed generation of O-allylic carbamates from amines, CO2, and allylic chlorides.

formation of a (π-allyl)palladium(II) intermediate, which ﬁxes generated CO2 through the amino-moiety. The resulting carbamate anion attacks the π-coordinated allyl group intramolecularly to afford the cyclic carbamate. In the presence of molecular iodine, the C–C double bond of allylamines and homoallylamines can be activated towards nucleophilic attack by carbamate anion. Accordingly, iodoalkyloxazolidinones and iodoalkyloxazinanones have been prepared under mild conditions (ambient temperature, atmospheric CO2 pressure) by the reaction of allylamines and homoallylamines with CO2 and iodine via intramolecular cyclization [88].

6.3 Synthesis of Carbamate Esters

6.3.6 Transfer of Carbamate Group to C–C Triple Bonds

The electrophilic activation of alkynes towards nucleophilic adducts of CO2 with amines allows the catalytic incorporation of the heterocumulene into enol carbamates. Sasaki and Dixneuf ﬁrst reported the synthesis of vinyl carbamates, in low yield, from diethylamine, 1-alkynes and CO2 (5 MPa, 398–413 K, 20 h) using Ru3(CO)12 as catalyst [89, 90c]. Later, several mononuclear Ru complexes were shown to be better catalysts for the activation of terminal alkynes towards ammonium N,N-dialkyl carbamates [90]. The formation of vinyl carbamates was, in general, highly regioselective with mononuclear complexes as compared to Ru3(CO)12, as the carbamate addition took place mainly at the unsubstituted carbon atom leading to the Z and E isomers. The Z-isomer was always more abundant than the E-isomer, and alkyne oligomers or polymers (polyphenylacetylene) were formed as side products. No carbamation occurred with disubstituted acetylenes such as diphenylacetylene, even under drastic conditions. The reaction was usually carried out in solvents such as diethylether, THF, acetonitrile, and toluene, but did not proceed in halogenated solvents. Recently, this process was reinvestigated using scCO2 as the reactant and solvent [22f, 91]. A comparative study showed that the synthesis of β-[(N,Ndiethylcarbamoyl)oxy]styrene from phenylacetylene and diethylamine in scCO2 was greatly accelerated for a series of ruthenium catalysts, compared to the same reactions in toluene [91]. In another recent report, the CO2-soluble ruthenium complex trans-[RuCl2{P(OC2H5)3}4] was found to promote, in scCO2, the carbamation reaction very selectively with respect to the side production of enynes obtained by dimerization of the alkyne substrates, affording Z-alkenyl carbamates with a very high stereoselectivity [22f]. While Ru3(CO)12 showed a very low activity towards the carbamation of acetylene, mononuclear ruthenium(II) complexes such as [RuCl2(η6-arene)]2 and RuX2PR3(η6-arene) (X = Cl, I; PR3 = PMe3, PBu3, PPh3, PMe2Ph) satisfactorily catalyzed the addition of N,N-dialkylcarbamates to HC≡CH [90c]. However, the best catalyst precursors were found to be RuCl3·xH2O [92] and [RuCl2(norbornadiene)]n [90c]. Alkyl primary amines did not undergo this reaction. The polymerization of acetylene into low and higher oligomers, catalyzed by ruthenium precursors, was noted and a threefold excess of acetylene with respect to the amine was required. Moreover, in addition to the primary product of [(N,N-dialkylcarbamoyl)oxy]ethylene, minor amounts of 2-[(N,N-dialkylcarbamoyl)oxy]buta-1,3-diene were also formed (Scheme 6.20). The latter species corresponds actually to the formal addition of carbamic acid to 3-buten-1-yne, the dimerization product of acetylene itself, which was formed as a side product during the process. Surprisingly, RuX2PR3(η6-arene) complexes did not promote the addition of ammonium N,N-dialkyl carbamates to alkenylacetylenes. However, this reaction was catalyzed by π-allyl ruthenium derivatives such as [Ph(CH2)nPPh2]Ru(η3CH2=C(Me)CH2)2 (n = 1–4), and yielded O-1-(1,3-dienyl)carbamates (4–62% yield)

145

146

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

Ru

R2NH + CO2 + HC CH

R2NCO2CH=CH2 + R2NCO2

C CH

CH2

CH2 R2N = Et2N,

N,

N, O

N

Scheme 6.20 Reaction of acetylene with secondary amines and CO2 in the presence of Ru-catalysts.

H C

Ru

Ru=C=CHR

C R

HC CR Ru-L

O R'2N

R'2NH2

L

O

O

O

O C NR'2

O C NR'2

Ru-C H

Ru-C CHR

+

H

R'2NH2+

CHR H+

Ru O O

O C NR'2 Ru-C

O C NR'2 H-C

O

CHR Ru

CH2R

O C NR'2 H-C Ru

CHR

Scheme 6.21 Carbamation of terminal alkynes promoted by Ru-catalysts: proposed mechanism.

regioselectively from CO2, secondary amines and isopropenylacetylene (373 K, 5 MPa CO2 pressure, 20 h) [93]. The dienyl carbamate esters corresponded to the addition of carbamate to the terminal carbon of triple bond. Scheme 6.21 shows the proposed mechanism for this class of transformations, which involve: (i) the formation of a ruthenium–vinylidene intermediate through

6.3 Synthesis of Carbamate Esters

the (η2-alkyne)-metal → (η1-alkylidene)-metal rearrangement; followed by (ii) a carbamate attack to the metal-bonded electrophilic carbon of the vinylidene– ruthenium moiety [90c]. To date, the activation of terminal alkynes HC≡CR′ (R′ = H, alkyl, aryl) towards the system R2NH/CO2 by catalysts based on transition metals other than Ru has been poorly explored. The only well-documented study has been that of Jiang and Hua, who investigated the catalytic activity of the rhenium complex ReBr(CO)5, which catalyzed the addition of Et2NH and CO2 (5.0 MPa) to terminal alkynes, affording anti-Markovnikov adducts of alkenyl carbamates (383 K, 24 h) in good to excellent yield and with high regioselectivity [94]. The reaction of CO2 and secondary aliphatic amines with α-ethynyl alcohols affords, in one step, O-β-oxoalkyl-N,N-dialkyl carbamates (Equation 6.12). R 2NH + CO2 + HC ≡ C − C (R1 )(R 2 ) OH → R 2NCO2C (R1 )(R 2 ) C (O) CH3

(6.12)

The process is catalyzed by ruthenium complexes such as [RuCl2(norbornadiene)]n, RuCl2(PMe2Ph)(p-cymene), Ru3(CO)12 (343–353 K, 5 MPa CO2 pressure, 20 h) [95], lanthanide metal chlorides (yields up to 38%) [96a], 1,1-bis(diphenylphosphino) ferrocene iron-carbonyl catalysts (yields up to 66%) [96b], and a cationic copper(I) complex based on a ferrocene-containing tetraazamacrocyclic ligand in the presence of bipyridine as base (yields 35–100%) [96c]. Recently, the efﬁcient synthesis of β-oxopropylcarbamates (up to 88% yield) via the three-component coupling of CO2, secondary amines and propargyl alcohols, was achieved in scCO2 in the absence of any additional catalyst and solvent (403 K, 14 MPa CO2 pressure, 48 h) [97]. Primary amines react with propargylic alcohols and CO2 in a different way, such that the reaction may afford cyclic carbamates. The reaction of n-propylamine with CO2 and HC≡CCH(R)OH (R = H, Me) in the presence of Ru3(CO)12 gave 3-propyl 2-oxo-1,3-oxazolines in 13–28% yield (353–373 K, 5 MPa CO2 pressure, 20 h) [95b]. Aliphatic or aromatic primary amines can react with CO2 and propargyl alcohols to afford also 4-methylene 3-substituted oxazolidin-2-ones; the latter process has been shown to be catalyzed by Cu(I) derivatives [98], or by a tertiary phosphine such as tri-n-butylphosphine (38–72% yield; 383–413 K, 5 MPa CO2, 20 h) [99]. The carbonylation of aliphatic primary amines to N-alkyl 4-methylene-2oxazolidinones by reaction with propargylic alcohols using CO2 as a carbonyl source proceeds efﬁciently with high yields (up to 95%) in ILs (393 K, 5 MPa CO2 pressure, 10 h) [100a]. An effective system, capable of working under less severe conditions (373 K, 2.5 MPa, 10–20 h), was developed by Deng and coworkers [100b], who used BMImBF4 as the solvent and CuCl as catalyst. Copper(I) halides can promote the cycloaddition reaction of CO2 with propargylic alcohols and aliphatic primary amines also in scCO2 (333 K, 8 MPa) to give 4-methylene-2-oxazolidinones or 4-methyloxazol-2-ones; under the above conditions, CuI was the catalyst of choice [100c]. The regiochemical control was seen to depend on the substituents of the propargylic alcohols, with tertiary propargylic

147

148

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

alcohols giving 4-methylene-2-oxazolidinones, and primary and secondary alcohols generating 4-methyloxazol-2(3H)-ones. The reaction of propargylamines with CO2 opens the way to the synthesis of 5-methylene-2-oxazolidinones. An early patent reported that these compounds could be prepared from acetylenic N-substituted amines and the heterocumulene, under copper catalysis, in yields up to 65% [101]. Likewise, (η4-1,5-cyclooctadiene) (η6-1,3,5-cyclooctatriene)ruthenium, Ru(COD)(COT), in the presence of PPh3, promoted the formation of 5-methylene N-alkyl oxazolidinones in good yields (63–80%; 373 K, 5 MPa CO2, 8 h) [90b]. This process can also be catalyzed by Pd-derivatives [102]. Shi and Shen have shown that Pd(O2CCH3)2 (5 mol%) also effectively catalyzed (323 K, 4 MPa, 48 h) the formation of N-unsubstituted 5-methylene-2-oxazolidinones from CO2 and N-unsubstituted propargylamines. The corresponding cyclic carbamates were formed as sole products, with good yield (50–85%). Unfortunately, internal alkynes, such as PhC≡CCH2NH2, reacted in a completely unselective manner. The carboxylation of N-alkylprop-2-ynylamines to 5-methylene-1,3-oxazolidin-2ones has been achieved, in good yields, using organic superbases having pKBH+ > 24 as catalysts, and acetonitrile as solvent, even under very mild conditions [103]. According to another procedure, the strong base can be generated electrochemically by the direct electrolysis of a solution of CH3CN and Et4NPF6 containing the amine. Subsequent CO2 bubbling (30 min) at ambient temperature and heating to reﬂux for 2 h gave the 5-methylene-1,3-oxazolidin-2-one in good yield (49–100%) [104]. Ikariya and coworkers ﬁrst investigated the carboxylative cyclization of N-alkyl propargylamines to 5-methylene-2-oxazolidinones in scCO2 [105a]. Under the working conditions (373 K, 9 MPa) the formation of cyclic urethanes proceeded in moderate to good yield in the absence of any catalyst. A subsequent report by a different group has shown that, in scCO2, the conversion of propargylamines into 5-methylene-2-oxazolidinones can be efﬁciently promoted by solid bases, such as basic alumina, hydrotalcites (MG30, MG70), or also organic bases supported on silica [105b].

6.4 Synthesis of Isocyanates

Isocyanates (RNCO) are industrially relevant compounds which ﬁnd application in several ﬁelds [9, 106]. Many isocyanates serve as the starting materials for the manufacture of plant protection agents, pesticides, dyes, resins and plastics, textile waterprooﬁng agents, detergents, bleaches, and adhesives. They are also widely used in surface coatings such as paints, sealants and ﬁnishes, and in the manufacture of rubbery plastics such as those used to coat wires. Traditionally, diisocyanates are the primary feedstock for the production of polyurethanes. The global market for diisocyanates in the year 2000 was 4.4 million tonnes, of which 61.3% was methylene diphenyl diisocyanate (MDI), 34.1% was toluene diisocyanate

6.4 Synthesis of Isocyanates

(TDI), 3.4% was the total for hexamethylene diisocyanate (HDI) and isophorone diisocyanate (IPDI), and 1.2% was the total for various others [106b]. Isocyanates may be prepared via a number of routes, including the thermal cracking of urethanes or silylurethanes [107], or the reaction of N-alkylcarbamate esters with boron trichloride [108] or chloroboranes [109] or chorosilanes [110] in the presence of bases (NEt3). To date, however, only amine phosgenation is practiced on a signiﬁcant industrial scale [10, 106]. The direct synthesis of isocyanates from CO2 is a challenging task, as the building up of the isocyanate group from CO2 requires deoxygenation of the heterocumulene. A number of metal systems are capable of promoting such a transformation in their coordination sphere. In 1976, Saegusa reported the copper(I) t-butoxide-promoted deoxygenation of CO2 by t-butyl isocyanide to produce t-butyl isocyanate and CO (Equation 6.13) [111a]. RNC + CO2 → RNCO + CO

(6.13)

This reaction can take place in mesitylene or tetralin at high temperature (373–423 K). Under the correct conditions (t-BuOCu/t-BuCN = 1 : 5 mol/mol; CO2/Cu = ca. 82 mol/mol; tetralin as solvent; 393 K, 3 h), the reaction produced up to 119% CO (based on copper) and an equivalent amount of isocyanate which, upon the addition of n-butyl alcohol, was converted in situ into (t-Bu)NHCO2Bu carbamate. An interesting question pertaining to the mechanism of Equation 6.13 is whether the CO carbon atom derives from CO2 or from t-BuNC. Unfortunately, the study did not provide any information on the mechanistic details of the process. An analogous reaction has been observed with the Ni(0) complex Ni(CNR)4 (R = Me; 2,6-Me2C6H3) which, in the presence of Li+ ions acting as catalysts of the process, reacted with CO2 to give Ni(CO)2(CNR)2 and RNCO [111b, c]. By means of isotope labeling studies, it was found that the pathway through which the CO ligands of Ni(CO)2(CNR)2 are formed involves multiple bond metathesis. In fact, the reaction of Ni(CNR)4 with 13CO2 yielded the unlabeled Ni–carbonyl complex, besides RN13CO. Moreover, a second labeling study using Ni(13CNR)4 and unlabeled CO2 afforded Ni(13CO)2(13CNR)2, together with unlabeled isocyanate. These results established, unambiguously, that the carbon atoms of the CO ligands in the complex Ni(CO)2(CNR)2 were derived from the isocyanide ligands of Ni(CNR)4, and not from CO2. Scheme 6.22 summarizes the suggested reaction pathway, which also accounts for the catalytic role of Li+. It is assumed that the Li+ ions may: (i) stabilize the development of negative charge on the oxygen atoms of CO2 ﬁxed by the N-atom of a coordinated isocyanide; and (ii) promote the interchange between a CO2 O-atom and an isocyanide RN-group. A strictly related example of multiple bond metathesis chemistry between CO2 and isocyanides proceeds from the binuclear complex Ni2(μ-CNMe)(CNMe)2(dppm)2 (dppm = 1,2-bis(diphenylphosphino)methane) which contains a reactive μ-MeNC ligand [111d–f]. This species, in the presence of liquid CO2 (298 K, 10–15 MPa, >48 h), was converted into Ni2(μ-CO)(CO)2(dppm)2 and polymeric

149

150

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

R

R Li+

* Ni=C=N

* + Ni=C=N O

C O

Li+

O

C O R N Ni

* C

C O

O

+

Li

-Li+

* Ni=C=O + O=C=NR

Scheme 6.22 Formation of organic isocyanates by alkali metal ion-catalyzed multiple bond metathesis reaction of CO2.

N,N′-dimethylcarbodiimide. It is believed that also in this case the initial fate of the CO2 carbon atom is very likely the formation of methyl isocyanate which, in this case, unfortunately could not be isolated as it further reacted to give carbodiimide. Isocyanate formation through multiple bond metathesis of CO2 with carbodiimide has been also demonstrated [112]. This transformation can be promoted by titanium isopropoxide, at 383 K, in THF as solvent. It is worth noting that the reverse process, which opens an entry into carbodiimide synthesis, is a well-known process that is catalyzed by several other systems, including trialkylphosphine oxides [113] or vanadium-oxo or -imido complexes [114]. The metathetical reaction of CO2 with metal (Ti(IV), U(V)) arylimido-complexes led to an arylisocyanate and a metal–oxo species [4g, 115a]. In these processes, which take place under very mild conditions (ambient temperature and pressure), the metal center acts as the sink for unrequired oxygen. A common feature which characterizes these transformations from a mechanistic point of view is the intermediate formation of a metal carbimato-species through the [2+2] cycloaddition of CO2 with the metal–imido complex (Scheme 6.23). By cycloreversion, the fourmembered aza-metalla-cycle then converts into isocyanate and a stable metal–oxo complex. These processes, at least formally, are reminiscent of the reaction of imino-phosphoranes with CO2 to give isocyanates and carbodiimides [115b,c]. Oxo-transfer from CO2 to a silyl group of bis-silyl-amido ligands provides an entry into silyl isocyanates or 1,3-bis(silyl) carbodiimides. Early studies have described this reaction for systems such as NaN(SiMe3)2 [116a] or Ln{N(SiMe3)2}3 (Ln = Pr, Nd) [116b]. Later, Sita and coworkers reported that this process could be promoted with a higher yield and selectivity by divalent Group 14 (Ge, Sn) bisamides (Scheme 6.24) [117a–c]. Closely related transformations have also been reported for Ti(IV) [117d] and Zr(IV) [117e].

6.4 Synthesis of Isocyanates

metal imido R

metal carbimato O

isocyanate formation O

C

C

O

R N M

N

C

M

O

R

N

R

O

N

O M

M

metal oxo "MO'' O=C=N-R

Scheme 6.23 Proposed mechanism for isocyanate formation via [2+2] cycloaddition of CO2 with a metal–imido complex.

SiMe3

SiMe3

2

Me3Si Me3Si

N

CO2 (0.4 MPa) M

N

pentane / 298 K

N Me3SiO-M

M-OSiMe3 + N

Me3SiN=C=O Me3SiN=C=NSiMe3

SiMe3

SiMe3 M = Ge, Sn

Scheme 6.24 Reaction of CO2 with divalent Group 14 bisamides.

SiMe3 Me3Si Me3Si

N

CO2 M

N SiMe3

SiMe3

O X M

Me3SiN=C=O

N O

+ SiMe3

M X

O

SiMe3

M = Ge, Sn Scheme 6.25 Metathetical exchange between CO2 and divalent Group 14 bisamides: proposed mechanism.

The mechanism through which the metathesis process is believed to occur with M[N(SiMe3)2]2 (M = Ge, Sn) amides is illustrated in Scheme 6.25 [117a]. The initial insertion step of the heterocumulene into the M–N bond is followed by a facile molecular rearrangement that involves the elimination of an isocyanate fragment. Due to the subvalent nature of the Group 14 metal center, it has been proposed that the carbamate ligand binds in a bidentate fashion. This particular mode of ligation might then facilitate the 1,3-shift of a trimethylsilyl group by the enhancement of its electrophilic character due to that of the carbonyl carbon. In a subsequent study, the same research group investigated the reactivity towards

151

152

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

PtBu2

Me2Si N

CO2

Ni t

SiMe2

P Bu2

PtBu2

Me2Si O Me2Si

Ni N

C O PtBu2

NBu4

NBu4+ S S

Ni N(SiMe3)2 S

+

S + CO2 -(Me3Si)2O

S

Ni N S

C O

Scheme 6.26 Deoxygenation of CO2 by sylylamido-ligands coordinated to Ni-centers: formation of isocyanate–Ni complexes.

CO2 of other monomeric stannilenes, such as bis[N-trimethylsilyl-N-2,6-diisopropylphenylamido]tin(II) and bis[N,N-bis(dimethylphenysilyl)amido]tin(II). Whilst the former stannilene was inert towards CO2, the latter one underwent metathesis (323 K, 0.2 MPa, pentane as solvent) with the heterocumulene generating the relevant isocyanate and carbodiimide in a 4.4 : 1 molar ratio [117b]. The full desilylation of silylamido-ligands has been observed with Ni-based systems (Scheme 6.26) [118]. These processes take place under mild conditions (0.1 MPa; 195 K for [Ni{N(SiMe3)2}‘S3’]– (‘S3’2– = (bis(2-mercaptophenyl)sulﬁde)2−); 213 K for (PNP)Ni (PNP = [(tBu2PCH2SiMe2)2N]1−), and result in the formation of a cyanate ligand which is N-bonded to the Ni center (isocyanato–metal complex). Both silyl groups of the silylamido moiety would act as oxygen acceptors, and show a pronounced tendency to migrate to the same CO2 oxygen. In the case of the nickel(II) complex [Ni{N(SiMe3)2}‘S3’]−, the wasted oxygen would be lost as Me3SiOSiMe3 [118a]. As for the nickel(I) compound (PNP)Ni, [(tBu2PCH2SiMe2)2O] NiNCO would be formed, as the product of transposition of the amide N and one CO2 oxygen [118b]. Remarkably, in the latter reaction the Ni(I)-complex reacted towards CO2 as an amide nucleophile more than as a reducing agent. Notably, the reaction was not dominated by any attempt from Ni(I) to achieve a higher coordination number or a higher oxidation state. Calderazzo has reported the synthesis of isocyanates from metal N-alkyl carbamates and acyl or aroyl chlorides (Scheme 6.27), used in this context as oxygen

6.4 Synthesis of Isocyanates M(O2CNHR)2 + 2 (RNH3)Cl

4 RNH2 + 2 CO2 + MCl2 M(O2CNHR)2 + 2MeC(O)Cl

(14a)

RNCO + MeC(O)OC(O)OMe + H2O + MCl2

4 RNH2 + 2 CO2 + 2 MeC(O)Cl

(14b)

RNCO + MeC(O)OC(O)OMe +2 (RNH3)Cl + H2O (14c)

Scheme 6.27 Metal-assisted synthesis of isocyanates from CO2, amines, and an acylating agent.

2 RNH2 + CO2

(RNH3)O2CNHR

RNHC(O)OH + iPrO2C-N-NCO2iPr

RNHC(O)OH + RNH2 RNCO + iPrO2C-NH-NHCO2iPr + R'3P=O

PR'3

Scheme 6.28 Synthesis of isocyanates from primary amines and CO2, using Mitsunobu chemistry

sinks (see also Section 6.3.2 and Scheme 6.6) [61a]. Isocyanates have been obtained in variable yields (2–42%), depending on several factors (nature of the metal, parent primary amine, acylating agent, solvent). Scheme 6.27 shows that the overall process corresponds to the metal-assisted synthesis of isocyanates from CO2, amines and an acylating agent. It is worth noting that the direct isocyanate synthesis through Equation 6.14c (Scheme 6.27) generates the target products in negligible yields (0–2%). The direct synthesis from amines and CO2 (Equation 6.15) is, in principle, a more appealing route to isocyanates, if suitable methods can be developed to drive to the right this thermodynamically unfavored [61a, 119] reaction: RNH2 + CO2 RNCO + H2O

(6.15)

A Mitsunobu-based procedure was developed by Jackson and coworkers for the preparation of alkyl and hindered aryl isocyanates in excellent yields from primary amines and CO2 under very mild conditions (Scheme 6.28) [120]. The protocol is based on the use of a Mitsunobu zwitterion generated from a dialkyl azodicarboxylate and PBu3 in dichloromethane at 195 K. Whilst the use of PPh3 still gave high yields of isocyanates from reactions with primary alkyl amines, only low yields were obtained from reactions with the aromatic amines. Those reactions which failed to produce high yields of isocyanates gave either carbamoylhydrazines and/ or dicarbamoylhydrazines and/or triazolinones as undesired side products. The use of dehydrating agents, in the presence of a tertiary organic co-base, also allows the high yield and selective synthesis of isocyanates, according to Equation 6.15 [121]. The research team at Monsanto has studied this process in depth, and have found in general the reaction to be fast and to require mild conditions (273 K, 0.1 MPa). Whilst at least 2 equiv. of the auxiliary base must be used, the co-base need not be a guanidine or phosphazene base in order to obtain a high yield (typically >90%) of isocyanate; more simply, triethylamine can be used. The dehydration agent is used in stoichiometric amounts with respect to the amine, and can

153

154

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

be selected among classical dehydration reagents, POCl3, PCl3, P4O10, SO3, and SOCl2. Acid anhydrides can also be used as dehydrating agents. Unfortunately, the use of large amounts of co-base, and the production of large quantities of wastes, makes the exploitation of this process difﬁcult on a large industrial scale.

6.5 Synthesis of Ureas

Interest in ureas [8] derives from their application in a wide variety of ﬁelds, as pharmaceuticals, petrochemicals, and agrochemicals. These compounds are used as dyes for cellulose ﬁbers, as antioxidants in gasoline, corrosion inhibitors, plant growth regulators, agroprotectives (pesticides and insecticides), and as tranquilizing and anticonvulsant agents. An unsymmetric ureidic group characterizes several biologically active compounds, with activity as inhibitors of HIV-1 protease or p38 kinase. Ureas are also useful intermediates in the production of carbamates [6, 7]. Apart from the synthesis of urea (NH2)2CO, the simplest homologue of this class of compounds and which is produced industrially from ammonia and CO2 by dehydration of the intermediate ammonium carbamate [119], the conventional method of synthesizing symmetrically N,N′-substituted ureas is based on the reaction of amines with phosgene. This approach is less efﬁcient for the synthesis of unsymmetrical ureas due to the important side formation of symmetrical ureas. The addition of amines to isocyanates provides the main route for the synthesis of unsymmetrical substituted ureas. During the past few years, harmful phosgene and isocyanates have been increasingly replaced with safer phosgene equivalents, such as triphosgene, activated carbonates, carbonyldiimidazole, carbamoyl chlorides, chloroformates, and also carbamates. However, most of these compounds are still usually prepared from phosgene as the starting material. McGhee et al., at Monsanto, have reported the synthesis of N,N-dialkylcarbamoyl chlorides from secondary amines and CO2 using thionyl chloride in the presence of an added tertiary amine base [122]. The use of CO2 as a starting material succedaneous for phosgene in the synthesis of substituted ureas represents an intriguing alternative to classical phosgenation methods, and several possible approaches have been explored to date. One such method is based on the decomposition of silylcarbamates [107h, 123k, l, n, o], which can be obtained from CO2 according to several methods [123]. In 1983, Knausz and coworkers reported that a primary trimethylsilyl carbamate RNHC(O)OSiMe3, when heated at temperatures between 373 and 393 K, converted into the corresponding urea (RNH)2CO according to the mechanism illustrated in Scheme 6.29 [123k]. Cyclic ureas could be obtained starting from trimethylsilyl esters of dicarbamic acids [107h]. On the basis of these results, Holmes and coworkers [123n] have developed a one-step synthesis of symmetric ureas (RNH)2CO (R = alkyl, aryl; 13–85% yield) from N-silylamines RNH(SiMe3) and scCO2 (393 K, 14 MPa, 17 h), used as solvent and reactant. The process involves the intermediate

6.5 Synthesis of Ureas

CO2

O 2 RNH O TMS

Δ

O RN O TMS

TMS +

RNH2

Δ O=C=N-R

O RHN

(TMS)2O

NHR

+ RNH2

Scheme 6.29 Synthesis of ureas by decomposition of O-trimethylsylyl carbamates.

formation of a silylcarbamate RNHC(O)OSiMe3, which then converts into the symmetric urea (RNH)2CO. It is worth noting that N-silyl derivatives of secondary amines RR ′N (SiR 3′′) , which react with scCO2 to give the corresponding carbamate, were found not to undergo the fragmentation reaction, even under forcing conditions; this fact supports the proposal for the involvement of a transient isocyanate intermediate (Scheme 6.29). The methodology has also been extended to the synthesis of unsymmetrical ureas RNHC(O)NR′R″. Trisubstituted ureas have been obtained, under mild conditions (333 K, 0.1 MPa, 6 h), in the presence of pyridine, from primary amines, CO2 and hexaalkylphosphorous triamides, P(NR2)3 [124]. The latter compounds are able to react with CO2 by converting into phosphocarbamate species, P(O2CNR2)3–x(NR2)x (x = 1, 2). The yields of ureas R2NC(O)NHR″ were almost quantitative, if based on one amino group of P(NR2)3, but could not be further improved by the addition of an excess of primary amine. This indicates that, in general, only one of the three P–N bonds of P(NR2)3 takes part in the reaction. The synthesis of ureas from amines and CO2 is a reversible process [125] which implies the elimination of water (Equation 6.16). 2 RR ′NH + CO2 → (RR ′N)2CO + H2O

(6.16)

Consequently, a variety of dehydrating agents, acting more properly as condensating reagents, has been investigated [126]; these include carbodiimides [126a, b], the commercially available adduct Me3N·SO3 [126c], phosphites [126d, e], phosphorous(III) chlorides [126f], and alkyl-triphenoxy-phosphonium halides [126g]. Each of these condensating reagents requires the use of an auxiliary base, which may be a tertiary amine, imidazole, a pyridine, or amidine base. Through these methods, symmetrical ureas (RNH)2CO (R = alkyl, aryl) can be obtained under very mild

155

156

NH2

R' R

CN

CO2 r.t.

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

R'

H N

C

H2O

NH R'

R

CNH2 O

O

O

O

NC

R

H2N

R C

r.t.

R'

H2O

N

NH

O HN

R

R'

H2N

R C

r.t.

R' N

NH

O O

R

R'

Scheme 6.30 Reaction of α-aminonitriles with CO2 at ambient temperature: synthesis of N-(3)-substituted hydantoins.

conditions, in moderate to excellent yield, depending on the amine and the auxiliary system (condensating agent/base) used. The C≡N functionality of α-aminonitriles may also act as a dehydrating/condensating center, as it may formally act as water trap and convert to an amide group [126h, i]. As shown in Scheme 6.30, at ambient temperature neat α-aminonitriles react with CO2 to afford unsymmetric disubstituted ureas which, in water and at room temperature, can be converted into N-(3)-substituted hydantoins. Symmetric ureas from primary amines and CO2 have been obtained through the modiﬁed “phosphine imide” reaction, using the system Ph3P/CCl4/NEt3 [76, 126j]. A Mitsunobu’s protocol, based on the use of triphenylphosphine and DEAD (diethylazodicarboxylate), has been developed for the synthesis, which provides good to excellent yields (80–98%) of symmetrically and unsymmetrically substituted ureas from a variety of primary and secondary amines using gaseous CO2 under mild conditions (room temperature, 0.1 MPa) [126k]. The above methodologies, although being intrinsically valuable from a strictly synthetic point of view, are of poor interest in terms of their application. Clearly, the direct synthesis of ureas from amines and the heterocumulene (Equation 6.16), without the use of auxiliary reagents, has a greater practical signiﬁcance. Acyclic ureas have been obtained in moderate yield (up to 66%, after 20 h) by the thermal decomposition of carbamate salts (RNH3)O2CNHR at high temperature (453 K), without the use of a catalyst [127a, b]. At high temperatures and pressures, diamines can also react with CO2, in the absence of a catalyst, to afford cyclic ureas (imidazolidin-2-ones, tetrahydro-pyrimidin-2-ones) in moderate to good yield and selectivity [69c, 127c]. The activity of some catalytic systems has been studied. For example, Deng has shown that the synthesis of symmetric ureas from aliphatic or aromatic primary amines and CO2 can be promoted by catalytic amounts of CsOH in several ILs [128a]. Under the working conditions employed (443 K, 6 MPa CO2 pressure), the best activity was shown by the system CsOH/BMImCl, whilst the urea yields (up to 98% within 4 h, for N,N′-dicyclohexylurea) were shown to depend on the amine used. Aromatic amines, such as aniline or p-methoxyaniline, were less reactive than their aliphatic counterparts and afforded the relevant urea in modest yield (27 and 33%, respectively), but only after long reaction times (36 h). In a later study, the same group showed that the synthesis of disubstituted symmetric ureas (RNH)2CO (R = alkyl) from amines and CO2 could be promoted also by catalytic

6.5 Synthesis of Ureas

amounts of a basic IL, such as 1-n-butyl-3-methyl imidazolium hydroxide ([BMIm]OH) [128b]. The group of De Vos has shown that Cs+ base catalysts (CsOH, CsF, Cs2CO3) can catalyze urea formation, when using N-methylpyrrolidone (NMP) as the solvent [125]. Under working conditions (443 K, 2.5 MPa), symmetrical urea, (RNH)2CO (R = alkyl) were prepared from CO2 and primary aliphatic amines in good yields (70–80%, after 24 h) and with high selectivity (100%), in most cases. Other organic solvents, such as n-decane, THF, CH3CN and DMF, proved to be less satisfactory reaction media, and gave poorer results than NMP. Compared to branched amines, the linear amines were considerably more reactive. Low yields (7%) were obtained from the sterically hindered primary amine t-BuNH2, and no urea compounds were prepared starting from secondary or aromatic amines. The use of a cocatalyst, such as Bu4NBr, further accelerated the formation of urea derivatives. In the presence of both a primary and a secondary amine, unsymmetrical ureas (RNH)C(O)(NR′R″) were obtained with variable yield and selectivity (67–95%). As secondary amines alone did not form urea compounds under the conditions used, the only side product was the symmetrical urea derived from the primary amine. The catalysts can be recovered and reused without any loss of activity, and the products were easily isolated from solution by precipitation with water. It has been reported that PdCl2(MeCN)2 catalyzed the synthesis of tetraethylurea from diethylamine and CO2, using PPh3/CCl4 as an auxiliary system in CH3CN [129a]. The reaction took place under mild conditions (room temperature, atmospheric pressure of CO2), but required a high concentration of PdCl2(MeCN)2 and a stoichiometric amount of triphenylphosphine, which underwent oxidation to phosphine oxide. The highest yield of urea was 7 TON (turnover number = moles of urea per mole Pd). Jessop, in a subsequent study, has reinvestigated this reaction and found that addition of the Pd-catalyst was unnecessary, as the phosphine alone had a weak ability to promote the process [129b]. Other bases, including 1,8-dimethylaminonaphthalene (DMAN), NEt2Ph, NPh3, and NEt3, were also tested for their ability to promote the reaction in the absence of Pd and PPh3. Although the promoting effect was moderate, the greater effect was found with DMAN. Other halide sources were less effective than CCl4 for urea synthesis. It has been speculated that the N,N-dialkylcarbamic acid may react with CCl4 to generate a CCl3-substituted intermediate that would be more reactive to nucleophilic attack by free amine. In the presence of molecular sieve 3A, triphenylstibine oxide (Ph3SbO) catalyzed the direct carbonylation of diamines H2N(CH2)nNHR (n = 2,3; R = Me, CH2CH2OH, CH2CHMeOH) with CO2 (5 MPa, 423–443 K), affording the corresponding cyclic ureas in high yield (>80%) [130a]. The modiﬁed catalytic system Ph3SbO/P4S10 was effective in promoting, under relatively milder conditions (353–423 K, 4.9 MPa, 12 h), the carbonylation of both mono- and di-amines to linear and cyclic ureas, respectively [130b, c]. Monitoring the reaction with 13C NMR spectroscopy revealed that the course of the reaction involved, as an intermediate step, the thiolation of initially formed carbamate salt to give an ionic carbamothiolate species which, upon aminolysis, converted into the urea product with the side production of H2S

157

158

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

BuNH3+ -O2CNHBu

2 BuNH2 + CO2

Ph3SbO/P4S10

(BuNH)2CO

Ph3SbO

O BuNH3

-H2S

NHBu S

Scheme 6.31 Carbonylation of amines with CO2 in the presence of the catalytic system Ph3SbO/P4S10.

+ Ru

H C

(RNH3)O2CNHR

H

Ru

C

O HO

OH

O C NHR

RNH2 HC C

OH O RHN C NHR

O OH

Ru

H

OH HO

Scheme 6.32 Ru-catalyzed direct synthesis of symmetrical N,N′-dialkylureas from CO2 and primary amines.

(Scheme 6.31). The Ph3SbO/P4S10 system also enabled the preparation of trisubstituted ureas by the cocarbonylation of primary and secondary aliphatic amines. In Section 6.3.6, it was emphasized that CO2 and secondary amines could add to terminal alkynes in the presence of ruthenium catalysts to afford carbamates. Under comparable conditions (393–413 K, 5 MPa; Ru-catalysts), primary amines will afford symmetrical disubstituted ureas in moderate yield [131]. It is worth noting that although the ﬁnal urea does not contain the starting alkyne, its catalytic formation requires, besides the Ru-catalyst, the presence of a stoichiometric amount of a 1-alkyne (e.g., a propargylic alcohol). A possible mechanism (Scheme 6.32) for this catalytic reaction may involve activation of the alkyne at the metal center, a nucleophilic addition of the carbamate to the activated alkyne to produce

References

a vinylcarbamate–ruthenium species, the addition of an amine to release urea and an enol–ruthenium intermediate. Thus, the ruthenium-alkyne system would behave as a dehydrating agent. Polymer-immobilized nanogold catalysts promote very efﬁciently the synthesis of symmetric disubstituted ureas (RNH)2CO (R = alkyl), with a turnover frequency (TOF) of up to 3000 h−1 [132]. After 20 h at 453 K, the reactions of CO2 (5 MPa) with cyclohexyl- or benzylamine gave the corresponding disubstituted ureas in 85% and 83% yields, respectively. A very recent study has shown that MCM-41 and HMS-type mesoporous silica are efﬁcient catalysts for the continuous chemical ﬁxation of CO2 into 1,3dimethyl-2-imidazolidinone with N,N-dimethylethylenediamine in scCO2 [133].

6.6 Conclusions

The ﬁxation of CO2 by amines or, more generally, by N-nucleophiles opens an entry into a large variety of N-carbonyl organic compounds through the synthesis of a new N–C bond. The relevant processes may play key roles in the development of a new “sustainable chemistry,” as in these reactions CO2 acts as a “green carbonyl source” in place of phosgene. During the past few decades, considerable progress has been made in this area, with catalytic processes having taken the place of stoichiometric reactions. Unfortunately, at present many of these processes are valuable only at the laboratory scale, with few of them appearing to have the potential for industrial exploitation. It is a fact that, since Bazarov’s discovery in 1870 [134], the synthesis of urea from ammonia and CO2 remains the sole industrially exploited example of CO2 ﬁxation into a new C–N bond. Yet, research in this area is far from being concluded. Notably, the use of scCO2 will undoubtedly open interesting future perspectives, including the quest for more eco-friendly reusable catalysts capable of promoting the ﬁxation of the heterocumulene with high yield and selectivity, under as much as possible mild conditions.

References 1 Aresta, M. and Quaranta, E. (1993) Proceedings of International Conference on Carbon Dioxide Utilization, Bari, Italy, pp. 63–77. 2 Belli Dell’Amico, D., Calderazzo, F., Labella, F., Marchetti, F., and Pampaloni, G. (2003) Chem. Rev., 103, 3857. 3 (a) Hill, M.R., Jensen, P., Russell, J.J., and Lamb, R.N. (2008) Dalton Trans., 2751.

(b) Notni, J., Schenck, S., Goerls, H., Breitzke, H., and Anders, E. (2008) Inorg. Chem., 47, 1382. (c) Baisch, U., Pagano, S., Zeuner, M., and Schnick, W. (2006) Eur. J. Inorg. Chem., 3517. (d) Verdejo, B., Aguilar, J., GarciaEspana, E., Gavina, P., Latorre, J., Soriano, C., Linares, J.M., and Domenech, A. (2006) Inorg. Chem., 45, 3803.

159

160

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas (e) Malik, M.A., O’Brien, P., Motevalli, M., and Abrahams, I. (2006) Polyhedron, 25, 241. (f) McCowan Shawn, C. and Caudle, M.T. (2005) Dalton Trans., 238. (g) Caudle, M.T., Brennessel, W.W., and Young, V.J. (2005) Inorg. Chem., 44, 3233. (h) Bramsen, F., Bond, A.D., McKenzie, C.J., Hazell, R.G., Moubaraki, B., and Murray, K.S. (2005) Chem. Eur. J., 11, 825. (i) Dell’Amico Belli, D., Calderazzo, F., Englert, U., Labella, L., Marchetti, F., and Specos, M. (2004) Eur. J. Inorg. Chem., 3938. (j) Aresta, M., Dibenedetto, A., and Quaranta, E. (1995) J. Chem. Soc. Dalton Trans., 3359. (k) Aresta, M. and Quaranta, E. (1995) J. Organomet. Chem., 488, 211. 4 (a) Tang, Y., Zackarov, L.N., Rheingold, A.L., and Kemp, R.A. (2004) Organometallics, 23, 4788. (b) Lill Nilsson, S.O., Kohn, U., and Anders, E. (2004) Eur. J. Org. Chem., 2868. (c) Tang, Y., Kassel, W.S., Zakharov, L.N., Rheingold, A.L., and Kemp, R.A. (2005) Inorg. Chem., 44, 359. (d) Mendiratta, A., Cummins, C.C., Cotton, F.A., Ibragimov, S.A., Murillo, C.A., and Villagran, D. (2006) Inorg. Chem., 45, 4328. (e) Himmel, H.J. (2007) Z. Anorg. Allg. Chem., 633, 2191. (f) Zeuner, M., Pagano, S., and Schnick, W. (2008) Chem. Eur. J., 14, 1524. (g) Bart, S.C., Anthon, C., Heinemann, F.W., Bill, E., Edelstein, N.M., and Meyer, K. (2008) J. Am. Chem. Soc., 130, 12536. 5 (a) Aresta, M. and Quaranta, E. (1988) J. Org. Chem., 53, 4153. (b) Aresta, M. and Quaranta, E. (1988) Italian Patent 1198206. (c) Aresta, M. and Quaranta, E. (1992) J. Chem. Soc. Dalton Trans., 1893. (d) Shi, M., Jiang, J.K., Shen, Y.M., Feng, Y.S., and Lei, G.X. (2000) J. Org. Chem., 65, 3443. (e) Shi, M., Shen, Y.M., and Jiang, J.K. (2001) Chin. J. Chem., 19, 405.

6 Adams, P. and Baron, F.A. (1965) Chem. Rev., 65, 567. 7 (a) Hagemann, C. (1983) Methoden der Organischen Chemie, 4th edn (ed. Houben-Weil), Georg Thieme Verlag, Stuttgart, vol. E4, p. 64. (b) Rossi, L. (2005) Sci. Synth., 18, 461. 8 (a) Gallou, I. (2007) Org. Prep. Proced. Intern., 39, 355. (b) Sartori, G. and Maggi, R. (2005) Sci. Synth., 18, 665. 9 (a) Braunstein, P.(1989) Chem. Rev., 89, 1927. (b) Twichett, H. (1974) Chem. Soc. Rev., 3, 209. 10 Cotarca, L. and Eckert, H. (2004) Phosgenations – A Handbook, Wiley-VCH Verlag GmbH, Weinheim. 11 (a) Aresta, M. and Quaranta, E. (1997) Chem. Tech., 27 (3), 32. (b) Carafa, M. and Quaranta, E. (2009) Mini-Rev. Org. Chem., 6, 168. 12 (a) Midvan, A.S., Fry, D.C., and Serpesu, E.H. (1990) Enzymatic and Model Carboxylations and Reactions for Carbon Dioxide Utilization (eds M. Aresta and J.V. Schloss), Kluwer, Dordrecht, p. 211. (b) Cleland, W.W., Andrews, T.J., Gutteridge, S., Hartmann, F.C., and Lorimer, G.H. (1998) Chem. Rev., 98, 549. 13 Schimming, V., Hoelger, C.G., Buntkowsky, G., Sack, I., Fuhrhop, J.H., Rocchetti, S., and Limbach, H.H. (1999) J. Am. Chem. Soc., 121, 4892. 14 Goettmann, F., Thomas, A., and Antonietti, M. (2007) Angew. Chem., Int. Ed., 46, 2717. 15 (a) Terlouw, J.K. and Schwarz, H. (1987) Angew. Chem., Int. Ed. Engl., 26, 805. (b) Khanna, R.K. and Moore, M.H. (1999) Spectrochim. Acta Part A, 55, 961. 16 Kaminskaia, N.V. and Kostic, N.M. (1997) Inorg. Chem., 36, 5917. 17 (a) Remko, M., Liedl, K.R., and Rode, B.M. (1993) J. Chem. Soc. Faraday Trans., 89, 2375. (b) Remko, M. and Rode, B.M. (1995) J. Mol. Struct. (THEOCHEM), 339, 125. (c) Wen, N. and Brooker, M.H. (1995) J. Phys. Chem., 99, 359.

References

18

19

20 21

22

23

24

25 26

(d) Jamroz, M.H., Dobrowolski, J.Cz., and Borowiak, M. (2000) Vib. Spectrosc., 22, 157. (e) Jamroz, M.H. and Dobrowolski, J. Cz. (2002) Vib. Spectrosc., 29, 217. Aresta, M., Ballivet Tkatchenko, D., Belli Dell’Amico, D., Bonnet, M.C., Boschi, D., Calderazzo, F., Faure, R., Labella, L., and Marchetti, F. (2000) Chem. Commun., 1099. Jamroz, M.H., Dobrowolski, J.Cz., Rode, J.E., and Borowiak, M. (2002) J. Mol. Struct. (THEOCHEM), 618, 101. Radeglia, R., Andersch, J., and Schroth, W. (1989) Z. Naturforsch., 44b, 181. (a) Masuda, K., Ito, Y., Horiguchi, M., and Fujita, H. (2005) Tetrahedron, 61, 213. (b) Hampe, E.M. and Rudkevich, D.M. (2003) Tetrahedron, 59, 9619. (a) Wittmann, K., Wisniewski, W., Mynott, R., Leitner, W., Kranermann, C.L., Rische, T., Eilbracht, P., Kluwer, S., Ernsting, J.M., and Elsevier, C.J. (2001) Chem. Eur. J., 7, 4584. (b) Furstner, A., Ackermann, L., Beck, K., Hori, H., Koch, D., Langemann, K., Liebl, M., Six, C., and Leitner, W. (2001) J. Am Chem. Soc., 123, 9000. (c) Xie, X., Liotta, C.L., and Eckert, C.A. (2004) Ind. Eng. Chem. Res., 43, 7907. (d) Fischer, H., Gyllenhaal, O., Vessmann, J., and Albert, K. (2003) Anal. Chem., 75, 622. (e) Dijkstra, Z.J., Doornbos, A.R., Weyten, H., Ernsting, J.M., Elsevier, C.J., and Keurentjes, J.T.F. (2007) J. Supercrit. Fluids, 41, 109. (f) Kayaki, Y., Suzuki, T., and Ikariya, T. (2008) Chem. Asian J., 3, 1865. (a) Aresta, M., Dibenedetto, A., Quaranta, E., Dibenedetto, A., Tommasi, I., and Marciniec, B. (2000) Appl. Organometal. Chem., 14, 871. (b) Dibenedetto, A., Aresta, M., Fragale, C., and Narracci, M. (2002) Green Chem., 4, 439. Alauzun, J., Besson, E., Mehdi, A., Reye, C., and Corriu, R.J.P. (2008) Chem. Mater., 20, 503. Smola, A. (1948) Mitt. Chem. Forsch. Inst. Ind. Oesterr., 2, 38. Neda, I., Kaukorat, T., and Fischer, A.K. (2003) Eur. J. Org. Chem., 3784.

27 (a) Aresta, M. and Quaranta, E. (1993) Italian Patent 1237208. (b) Aresta, M. and Quaranta, E. (1992) Tetrahedron, 48, 1515. 28 (a) Danckwerts, P.V. (1979) Chem. Eng. Sci., 34, 443. (b) Crooks, J.E. and Donnellan, J.P. (1989) J. Chem. Soc., Perkin Trans. II, 331. (c) Jamroz, M.H., Dobrowolski, J.Cz., and Borowiak, M. (1997) J. Mol. Struct., 404, 105. 29 (a) Hori, Y., Nagano, Y., Nakao, J., Fukuoka, T., and Taniguchi, H. (1986) Chem. Express, 1, 224. (b) McGhee, W.D., Riley, D., Kevin, C., Pan, Y., and Parnas, B. (1995) J. Org. Chem., 60, 2820. (c) McGhee, W.D., Pan, Y., and Riley, D.P. (1994) J. Chem. Soc. Chem. Commun., 699. (d) McGhee, W.D., Barry, L., Riley, D.P., and Talley, J.J. (1992) European Patent 511948. 30 (a) Kurz, F., Rumpf, B., Sing, R., and Maurer, G. (1996) Ind. Eng. Chem. Res., 35, 3795. (b) Buckingham, A.D., Handy, N.C., Rice, J.E., Somasundram, K., and Dijkgraaf, C. (1986) J. Comput. Chem., 7, 283. 31 (a) Mani, F., Peruzzini, M., and Stoppioni, P. (2006) Green Chem., 8, 995. (b) Filburn, T., Helble, J.J., and Weiss, R.A. (2005) Ind. Eng. Chem. Res., 44, 1542. (c) Bottinger, W., Maiwald, M., and Hasse, H. (2008) Fluid Phase Equilibria, 263, 131. 32 (a) Bates, E.D., Mayton, R.D., Ntai, I., and Davis, J.H., Jr (2002) J. Am. Chem. Soc., 124, 926. (b) Blanchard, L.A., Gu, Z., and Brennecke, J.F. (2001) J. Phys. Chem. B, 105, 2437. 33 Kovvali, A.S. and Sirkar, K.K. (2001) Ind. Eng. Chem. Res., 40, 2502. 34 (a) George, M. and Weiss, R.J. (2001) J. Am. Chem. Soc., 123, 10393. (b) Carretti, E., Dei, L., Baglioni, P., and Weiss, R.G. (2003) J. Am. Chem. Soc., 125, 5121.

161

162

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas 35 Stastny, V., Anderson, A., and Rudkevivh, D.M. (2006) J. Org. Chem., 71, 8696. 36 Ki, C.D., Oh, C., Oh, S.G., and Chang, J.Y. (2002) J. Am. Chem. Soc., 124, 14838. 37 (a) Yamada, T., Lukac, P.J., George, M., and Weiss, R.G. (2007) Chem. Mater., 19, 967. (b) Phan, L., Andreatta, J.R., Horvey, L.K., Edie, C.F., Luco, A.L., Mirchandani, A., Darensbourg, D.J., and Jessop, P.J. (2008) J. Org. Chem., 73, 127. 38 Ray, S., Pathak, S.R., and Chaturvedi, D. (2005) Drugs Fut., 30, 161. 39 Gupta, R.C. (2006) Toxicology of Organophosphate and Carbamate Compounds (ed. R.C. Gupta), Elsevier Inc., Burlington, p. 5. 40 (a) Green, T.W. and Wuts, P.G.M. (2007) Protective Groups in Organic Synthesis, John Wiley & Sons, Inc., New York. (b) Buchstaller, H.P. (1998) Tetrahedron, 54, 3465. (c) Ullrich, H. (1983) in Kirk-Othmer, Encyclopedia of Chemical Technology, vol. 23, 3rd edn, John Wiley & Sons Inc., New York, p. 576. 41 Paul, F. (2000) Coord. Chem. Rev., 203, 269. 42 (a) Giannoccaro, P., Dibenedetto, A., Gargano, M., Quaranta, E., and Aresta, M. (2008) Organometallics, 27, 967. (b) Giannoccaro, P., Cornacchia, D., Doronzo, S., Mesto, E., Quaranta, E., and Aresta, M. (2006) Organometallics, 25, 2872. 43 Delledonne, F., Rivetti, F., and Romano, U. (2001) Appl. Cat. A: Gen., 221, 241. 44 Chaturvedi, D. and Ray, S. (2007) Curr. Org. Chem., 11, 987. 45 Tsuda, T., Washita, H., Watanabe, K., Miwa, M., and Saegusa, T. (1978) J. Chem. Soc. Chem. Commun., 815. 46 Belforte, A. and Calderazzo, F. (1989) J. Chem. Soc., Dalton Trans., 1007. 47 Yoshida, Y., Ishii, S., Watanabe, M., and Yamashita, T. (1989) Bull. Chem. Soc. Jpn, 62, 1534. 48 (a) Shi, M. and Shen, Y.M. (2001) Helv. Chim. Acta, 84, 3357.

49 50 51

52 53 54

55

56

57

58

(b) Perez, E.R., Odnicki da Silva, M., Costa, V.C., Rodriguez-Filho, U.P., and Franco, D.W. (2002) Tetrahedron Lett., 43, 4091. McGhee, W.D. and Talley, J.J. (1994) US Patent 5302717. Butcher, K.J. (1994) Synlett, 825. (a) Salvatore, R.N., Flanders, V.L., Ha, D., and Jung, K.W. (2000) Org. Lett., 2, 2797. (b) Jung, K.W. (2000) WO 0050389. (c) Salvatore, R.N., Shin, S.I., Nagle, A.S., and Jung, K.W. (2001) J. Org. Chem., 66, 1035. (d) Salvatore, R.N., Ledger, J.A., and Jung, K.W. (2001) Tetrahedron Lett., 42, 6023. (e) Salvatore, R.N., Chu, F., Nagle, A.S., Kapxhiu, E.A., Cross, R.M., and Jung, K.W. (2002) Tetrahedron, 58, 3329. Yoshida, M., Hara, N., and Okuyama, S. (2000) Chem. Commun., 151. Singh, K.N. (2007) Synth. Commun., 37, 2651. (a) Chaturvedi, D., Mishra, N., and Mishra, V. (2006) Chinese Chem. Lett., 17, 1309. (b) Chaturvedi, D. and Ray, S. (2005) Lett. Org. Chem., 2, 742. (a) Chaturvedi, D. and Ray, S. (2006) Monatsh. Chem., 137, 201. (b) Chaturvedi, D. and Ray, S. (2006) Monatsh. Chem., 137, 459. Chaturvedi, D., Kumar, A., and Ray, S. (2002) Synth. Commun., 32, 2651. (a) Casadei, M.A., Micheletti Moracci, F., Zappia, G., Inesi, A., and Rossi, L. (1997) J. Org. Chem., 62, 6754. (b) Feroci, M., Inesi, A., and Rossi, L. (2000) Tetrahedron Lett., 41, 963. (c) Feroci, M., Casadei, M.A., Orsini, M., Palombi, L., and Inesi, A. (2003) J. Org. Chem., 68, 1548. (a) Hess, U., Komenda, J., Hollwarth, A., and Dunkel, S. (1992) Pharmazie, 47, 848. (b) Casadei, M.A., Inesi, A., Micheletti Moracci, F., and Rossi, L. (1996) Chem. Commun., 2575. (c) Feroci, M., Orsini, M., Rossi, L., Sotgiu, G., and Inesi, A. (2007) J. Org. Chem., 72, 200.

References

59

60

61

62

63

64

65

66

(d) Niu, D.F., Zhang, L., Xiao, L.P., Luo, Y.W., and Lu, J.X. (2007) Appl. Organomet. Chem., 21, 941. (a) Ikeda, S., Takagi, T., and Ito, K. (1987) Bull. Chem. Soc. Jpn, 60, 2517. (b) Calderwood, J.L., Jr, Roberts, T.S., and Sawyer, D.T. (1984) J. Am. Chem. Soc., 106, 4667. (a) Srivastava, R., Manju, M.D., Srinivas, D., and Ratnasamy, P. (2004) Catal. Lett., 97, 41. (b) Srivastava, R., Srinivas, D., and Ratnasamy, P. (2005) Appl. Catal. A: General, 289, 128. (c) Srivastava, R., Srinivas, D., and Ratnasamy, P. (2006) Tetrahedron Lett., 47, 4213. (d) Srivastava, R., Srinivas, D., and Ratnasamy, P. (2005) J. Catal., 233, 1. (e) Srivastava, R., Srinivas, D., and Ratnasamy, P. (2006) Micropor. Mesopor. Mater., 90, 314. (a) Belforte, A., Belli Dell’Amico, D., and Calderazzo, F. (1988) Chem. Ber., 121, 1891. (b) Belforte, A., Belli Dell’Amico, D., Calderazzo, F., Giurlani, U., and Labella, L. (1993) Gazz. Chim. Ital., 123, 119. (a) Aresta, M. and Quaranta, E. (1993) Italian Patent 1237207. (b) Aresta, M. and Quaranta, E. (1991) Tetrahedron, 47, 9489. (c) Aresta, M., Dibenedetto, A., Quaranta, E., Boscolo, M., and Larsson, R. (2001) J. Mol. Catal., 174, 7. (d) Aresta, M. and Dibenedetto, A. (2002) Chem. Eur. J., 8, 685. Ishii, S., Nakayama, H., Yoshida, Y., and Yamashita, T. (1989) Bull. Soc. Chem. Jpn, 62, 455. (a) Abla, M., Choi, J.C., and Sakakura, T. (2001) Chem. Commun., 2238. (b) Abla, M., Choi, J.C., and Sakakura, T. (2004) Green Chem., 6, 524. (c) Choi, J.C., Mahamdi, A., and Sakakura, T. (2006) Japanese Patent 022043. Ion, A., Van Doorslaer, C., Parvulescu, V., Jacobs, P., and De Vos, D. (2008) Green Chem., 10, 111. Bratt, M.O. and Taylor, P.C. (2003) J. Org. Chem., 68, 5439.

67 (a) Chaturvedi, D., Kumar, A., and Ray, S. (2003) Tetrahedron Lett., 44, 7637. (b) Chaturvedi, D., Mishra, N., and Mishra, V. (2007) Monatsh. Chem., 138, 57. 68 (a) Soga, K., Hosoda, S., Nakamura, H., and Ikeda, S. (1976) J. Chem. Soc. Chem. Commun., 617. (b) Kawanami, H. and Ikushima, Y. (2002) Tetrahedron Lett., 43, 3841. (c) Kawanami, H., Matsumoto, H., and Ikushima, Y. (2005) Chem. Lett., 34, 60. (d) Du, Y., Wu, Y., Liu, A.H., and He, L.N. (2008) J. Org. Chem., 73, 4709. (e) Nomura, R., Nakano, T., Nishio, Y., Ogawa, S., Ninagawa, A., and Matsuda, H. (1989) Chem. Ber., 122, 2407. (f) Shen, Y.M., Duan, W.L., and Shi, M. (2004) Eur. J. Org. Chem., 3080. (g) Hancock, M.T. and Pinhas, A.R. (2003) Tetrahedron Lett., 44, 5457. (h) Hancock, M.T. and Pinhas, A.R. (2004) Synthesis, 14, 2347. (i) Mu, W.H., Chasse, G.A., and Fang, D.C. (2008) J. Phys. Chem. A, 112, 6708. (j) Sudo, A., Morioka, Y., Koizumi, E., Sanda, F., and Endo, T. (2003) Tetrahedron Lett., 44, 7889. (k) Sudo, A., Morioka, Y., Sanda, F., and Endo, T. (2004) Tetrahedron Lett., 45, 1363. (l) Miller, A.W. and Nguyen, S.T. (2004) Org. Lett., 6, 2301. (m) Tascedda, P. and Dunach, E. (2000) Chem. Commun., 449. 69 (a) Steele, A.B. (1959) US Patent 2868801. (b) Lynn, J.W. (1961) US Patent 2975187. (c) Bhanage, B.M., Fujita, S., Ikushima, Y., and Arai, M. (2003) Green Chem., 5, 340. 70 (a) Matsuda, H., Baba, A., Nomura, R., Kori, M., and Ogawa, S. (1985) Ind. Eng. Chem. Prod. Res. Dev., 24, 239. (b) Tominaga, K. and Sasaki, Y. (2002) Synlett, 307. (c) Fujita, S., Kanamaru, H., Senboku, H., and Arai, M. (2006) Int. J. Mol. Sci., 7, 438. 71 Akasaki, Y., Hatano, S., and Fukuyama, M. (1978) Japanese Patent 78111062. 72 (a) Dozen, Y. and Ukita, M. (1982) Japanese Patent 57042678.

163

164

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas

73

74

75

76

77

78

79 80 81 82 83

(b) Kawanami, H. and Ikushima, Y.J. (2002) Jpn. Petrol. Inst., 45, 321. (a) Feroci, M., Inesi, A., Mucciante, V., and Rossi, L. (1999) Tetrahedron Lett., 40, 6059. (b) Casadei, M.A., Feroci, M., Inesi, A., Rossi, L., and Sotgiu, G. (2000) J. Org. Chem., 65, 4759. (c) Feroci, M., Gennaro, A., Inesi, A., Orsini, M., and Palombi, L. (2002) Tetrahedron Lett., 43, 5863. (a) Kodaka, M., Tomohiro, T., Lee, A.L., and Okuno, H. (1989) J. Chem. Soc. Chem. Commun., 1479. (b) Kodaka, M., Lee, A.L., Tomohiro, T., and Okuno, H. (1990) Chem. Express, 5, 233. (c) Kodaka, M., Tomohiro, T., Lee, A.L., and Okuno, H. (1991) Chem. Express, 6, 217. (d) Ariza, X., Pineda, O., Urpì, F., and Villarasa, J. (2001) Tetrahedron Lett., 42, 4995. (a) Kodaka, M., Tomohiro, T., and Oxuno, H.H. (1993) J. Chem. Soc. Chem. Commun., 81. (b) Dinsmore, C.J. and Mercer, S.P. (2004) Org. Lett., 6, 2885. Kubota, Y., Kodaka, M., Tomohiro, T., and Okuno, H. (1993) J. Chem. Soc., Perkin Trans. I, 5. (a) Yoshida, Y., and Inoue, S. (1978) Bull. Chem. Soc. Jpn, 51, 559. (b) Yoshida, Y. and Inoue, S. (1980) Polym. J., 12, 763. (c) Yoshida, Y., Ishii, S., Kawato, A., Yamashita, T., Yano, M., and Inoue, S. (1988) Bull. Chem. Soc. Jpn, 61, 2913. (a) Yoshida, Y. and Inoue, S. (1978) Chem. Lett., 139. (b) Yoshida, Y. and Inoue, S. (1979) J. Chem. Soc., Perkin Trans. I, 3146. Asano, T., Saito, N., Ito, S., Hatakeda, K., and Toda, T. (1978) Chem. Lett., 311. Kojima, F., Aida, T., and Inoue, S. (1986) J. Am. Chem. Soc., 106, 391. Yoshida, M., Ohshima, M., and Toda, T. (1993) Heterocycles, 35, 623. Karikomi, M., Yamazaki, T., and Toda, T. (1993) Chem. Lett., 1965. (a) Toda, T. (1977) Chem. Lett., 957. (b) Saito, N., Hatakeda, K., Ito, S., Asano, T., and Toda, T. (1986) Bull. Chem. Soc. Jpn, 59, 1629.

84 Ishii, S., Zhou, M., Yoshida, Y., and Noguchi, H. (1999) Synth. Commun., 29, 3207. 85 Yoshida, Y. and Inoue, S. (1977) Chem. Lett., 1375. 86 McGhee, W.D. and Riley, D.P. (1992) Orgamometallics, 11, 900. 87 (a) McGhee, W.D., Riley, D.P., Christ, M.E., and Christ, K.M. (1993) Orgamometallics, 12, 1429. (b) Yoshida, M., Ohsawa, Y., Sugimoto, K., Tokuyama, H., and Masataka, I. (2007) Tetrahedron Lett., 48, 8678. 88 Toda, T. and Kitagawa, Y. (1987) Angew. Chem., Int. Ed. Engl., 26, 334. 89 Sasaki, D. and Dixneuf, P.H. (1986) J. Chem. Soc. Chem. Commun., 790. 90 (a) Mahé, R., Dixneuf, P.H., and Lécolier, S. (1986) Tetrahedron Lett., 27, 6333. (b) Mitsudo, T., Hori, Y., Yamakawa, Y., and Watanabe, Y. (1987) Tetrahedron Lett., 28, 4417. (c) Mahé, R., Sasaki, Y., Bruneau, C., and Dixneuf, P.H. (1989) J. Org. Chem., 54, 1518. (d) Bruneau, C. and Dixneuf, P.H. (1992) J. Mol. Catal., 74, 97. 91 Rohr, M., Geyer, C., Wandeler, R., Schneider, M.S., Murphy, E.F., and Baiker, A. (2001) Green Chem., 3, 123. 92 Sasaki, Y. and Dixneuf, P.H. (1987) J. Org. Chem., 52, 314. 93 Hofer, J., Doucet, H., Bruneau, C., and Dixneuf, P.H. (1991) Tetrahedron Lett., 32, 7409. 94 Jiang, J.L. and Hua, R. (2006) Tetrahedron Lett., 47, 953. 95 (a) Bruneau, C. and Dixneuf, P.H. (1987) Tetrahedron Lett., 28, 2005. (b) Sasaki, Y. and Dixneuf, P.H. (1987) J. Org. Chem., 52, 4389. 96 (a) Shim, S.C., Baeg, J.O., Doh, C.H., Youn, Y.Z., and Kim, T.J. (1990) Bull. Korean Chem. Soc., 11, 467. (b) Kim, T.J., Kwon, K.H., Kwon, S.C., Baeg, J.O., and Shim, S.C. (1990) J. Organomet. Chem., 389, 205. (c) Kim, H.S., Kim, J.W., Kwon, S.C., Shim, S.C., and Kim, T.J. (1997) J. Organomet. Chem., 545-546, 337. 97 Qi, C.R. and Jiang, H.F. (2007) Green Chem., 9, 1284.

References 98 Dimroth, P., Pasedach, H., and Schefczik, E. (1964) German Patent 1151507. 99 Fournier, J., Bruneau, C., and Dixneuf, P.H. (1990) Tetrahedron Lett., 31, 1721. 100 (a) Zhang, Q., Shi, F., Gu, Y., Yang, J., and Deng, Y. (2005) Tetrahedron Lett., 46, 5907. (b) Gu, Y., Zhang, Q., Duan, Z., Zhang, J., Zhang, S., and Deng, Y. (2005) J. Org. Chem., 70, 7376. (c) Jiang, H., Zhao, J., and Wang, A. (2008) Synthesis, 763. 101 Dimroth, P. and Pasedach, H. (1964) German Patent 1164411. 102 Shi, M. and Shen, Y.M. (2002) J. Org. Chem., 67, 16. 103 Costa, M., Chiusoli, G.P., Taffurelli, D., and Dalmonego, G. (1998) J. Chem. Soc., Perkin Trans. I, 1541. 104 Feroci, M., Orsini, M., Sotgiu, G., Rossi, L., and Inesi, A. (2005) J. Org. Chem., 70, 7795. 105 (a) Kayaki, Y., Yamamoto, M., Suzuki, T., and Ikariya, T. (2006) Green Chem., 8, 1019. (b) Maggi, R., Bertolotti, C., Orlandini, E., Oro, C., Sartori, G., and Selva, M. (2007) Tetrahedron Lett., 48, 2131. 106 (a) Oertel, G.(1985) Polyurethane Handbook, Hanser Publishers, Munich. (b) Randall, D. and Lee, S. (2002) The Polyurethane Book, John Wiley & Sons, Inc.. 107 (a) Merger, F. and Towae, F. (1987) US Patent 4713476. (b) Sundermann, R., Konig, K., Engbert, T., Becker, G., and Hammen, G. (1983) US Patent 4388246. (c) Rosenthal, R. and Zajacek, J.G. (1976) US Patent 3962302. (d) Bortnick, N., Luskin, L.S., Hurwitz, M.D., and Rytina, A.W. (1956) J. Am. Chem. Soc., 78, 4358. (e) Spohn, R.J. (1983) GB Patent 2113673. (f) Mironov, V.F., Kozyukov, V.P., and Bulatov, V.P. (1973) Zh. Obshch. Khim., 43, 2089. (g) Kozyukov, V.P., Orlov, G.I., and Mironov, V.F. (1979) Zh. Obshch. Khim., 49, 2155. (h) Szalay, R., Bocskei, Z., Knausz, D., Lovasz, C., Ujszaszy, K., Szakacz, I.,

108 109 110

111

112 113 114

115

116

and Sohar, P. (1996) J. Organomet. Chem., 510, 93. Alper, H. and Butler, D.C.D. (1998) Chem. Commun., 2575. Alper, H. and Valli, V.L.K. (1995) J. Org. Chem., 60, 257. (a) Chong, P.J., Janicki, S.Z., and Petillo, P.A. (1998) J. Org. Chem., 63, 8515. (b) Kricheldorf, H.R. (1970) Synthesis, 649. (c) Pirkle, W.H. and Rinaldi, P.L. (1978) J. Org. Chem., 43, 3803. (a) Tsuda, T., Sanada, S.I., and Saegusa, T. (1976) J. Organomet. Chem., 116, C10. (b) Kim, W.Y., Chang, J.S., Park, S.E., Ferrence, G., and Kubiak, C.P. (1998) Chem. Lett., 1063. (c) Kim, W.Y., Chang, J.S., Park, S.E., and Kubiak, C.P. (1999) Res. Chem. Intermed., 25, 459. (d) DeLaet, D.L., Fanwick, P.E., and Kubiak, P.C. (1987) J. Chem. Soc. Chem. Commun., 1412. (e) DeLaet, D.L., Fanwick, R., del Rosario, P.E., and Kubiak, P.C. (1987) J. Am. Chem. Soc., 109, 754. (f) Kubiak, P.C., and Ratliff, K.S. (1991) Israel J. Chem., 31, 3. Ghosh, R. and Samuelson, A.G. (2005) Chem. Commun., 2017. William, A. and Ibrahim, I.T. (1981) Chem. Rev., 81, 589. (a) Birdwhistell, K.R., Boucher, T., Ensminger, M., Harris, S., Johnson, M., and Toporek, S. (1993) Organometallics, 12, 1023. (b) Birdwhistell, K.R., Lanza, J., and Pasos, J. (1999) J. Organomet. Chem., 584, 200. (a) Kilgore, U.J., Basuli, F., Huffmann, J.C., and Mindiola, D.J. (2006) Inorg. Chem., 45, 487. (b) Wadsworth, W.S. and Emmons, W.D. (1964) J. Org. Chem., 29, 2816. (c) Molina, P., Alajarin, M., and Arques, A. (1982) Synthesis, 596. (a) Wannagat, U., Kuckertz, H., Kruger, C., and Pump, J. (1964) Z. Anorg. Allg. Chem., 333, 54. (b) Radkov, Y.F., Fedorova, E.A., Khorshev, S.Y., Kalinina, G.S., Bochkarev, M.N., and Razuvaev, G.A. (1986) Zh. Obshch. Khim., 56, 386.

165

166

6 Synthesis of Carbamic Acids and Their Derivatives, Isocyanates, and Ureas 117 (a) Sita, L.R., Babcock, J.R., and Xi, R. (1996) J. Am. Chem. Soc., 118, 10912. (b) Babcock, J.R., Sands, L.L., Rheingold, A.L., and Sita, L.R. (1999) Organometallics, 18, 4437. (c) Babcock, J.R. and Sita, L.R. (1998) J. Am. Chem. Soc., 120, 5585. (d) Ciruelos, S., Cuenca, T., Gomez, R., Gomez-Sal, P., Manzanero, A., and Royo, P. (1996) Organometallics, 15, 5577. (e) Kloppenburg, L. and Petersen, J.L. (1996) Organometallics, 15, 7. 118 (a) Sellmann, D., Geipel, F., and Heinemann, F.W. (2000) Chem. Eur. J., 6, 4279. (b) Fullmer, B.C., Fan, H., Ink, M., and Caulton, K.G. (2008) Inorg. Chem., 47, 1865. 119 Tsipis, C.A. and Karipidis, P.A. (2005) J. Phys. Chem. A, 109, 8560. 120 (a) Horvath, M.J., Saylik, D., Elmes, P.S., Jackson, W.R., Lovel, C.G., and Moody, K. (1999) Tetrahedron Lett., 40, 363. (b) Saylik, D., Horvath, M.J., Elmes, P.S., Jackson, W.R., Lovel, C.G., and Moody, K. (1999) J. Org. Chem., 64, 3940. 121 (a) Riley, D., McGhee, W.D., and Waldmann, T. (1994) ACS Symp. Ser., 577 (Benign by Design), 122. (b) Waldmann, T.E. and McGhee, W.D. (1994) J. Chem. Soc. Chem. Commun., 957. (c) McGhee, W.D., Paster, M., Riley, D., Ruettimann, K., Solodar, J., and Waldmann, T. (1996) ACS Symp. Ser., 626 (Green Chemistry), 49. 122 McGhee, W.D., Pan, Y., and Talley, J.J. (1994) Tetrahedron Lett., 35, 839. 123 (a) Brederweld, H. (1962) Recl. Trav. Chim. Pays-Bas, 81, 276. (b) Birkofer, L. and Krebs, K. (1968) Tetrahedron Lett., 9, 885. (c) Mironov, V.F., Kozyukov, V.P., Kirilin, A.D., Sheludyakov, V.D., Dergunov, J.J., and Vostokov, L.A. (1975) Zh. Obshch. Khim., 45, 2007. (d) Oertel, G., Holtschmidt, H., and Malz, H. (1964) Chem. Ber., 97, 891. (e) Yoder, C.H., Komoriya, A., Kochanowski, J.E., and Suydam, F.H. (1971) J. Am. Chem. Soc., 93, 6515.

(f) Cragg, R.H. and Lappert, M.F. (1966) J. Chem. Soc. A, 82. (g) Ebsworth, E.A.V., Rocktaschel, G., and Thompson, J.A. (1967) J. Chem. Soc. A, 32. (h) Skeludyakov, V.D., Kirilin, A.D., and Mironov, V.F. (1975) Zh. Obshch. Khim., 45, 479. (i) Skeludyakov, V.D., Kirilin, A.D., Gusev, V.A., Sharapov, V.A., and Mironov, V.F. (1976) Zh. Obshch. Khim., 46, 2712. (j) Birkofer, L. and Sommer, P. (1972) J. Organomet. Chem., 35, C15. (k) Knausz, D., Meszticcky, L., Szakacs, B., Csakvari, B., and Ujszaszy, K. (1983) J. Organomet. Chem., 256, 11. (l) Zoeckler, M.T. and Laine, M.R. (1983) J. Org. Chem., 43, 2539. (m) Lipshutz, B.H., Papa, P., and Keith, J.M. (1999) J. Org. Chem., 64, 3792. (n) Fuchter, M.J., Smith, C.J., Tsang, M.W.S., Boyer, A., Saubern, S., Ryan, J.H., and Holmes, A.B. (2008) Chem. Commun., 2152. (o) Jimenez, G., Rodriguez, E., Gomez Sal, P., Royo, P., Cuenca, T., and Galakhov, M. (2001) Organometallics, 20, 2459. 124 Yamazaki, N., Tomioka, T., and Higashi, F. (1975) Synthesis, 384. 125 Ion, A., Parvulescu, V., Jacobs, P., and De Vos, D. (2007) Green Chem., 9, 158. 126 (a) Ogura, H., Takeda, K., Tokue, R., and Kobayashi, T. (1978) Synthesis, 394. (b) Aresta, M., Dibenedetto, A., Fracchiolla, E., Giannoccaro, P., Pastore, C., Papai, I., and Schubert, G. (2005) J. Org. Chem., 70, 6177. (c) Cooper, C.F. and Falcone, S.J. (1995) Synth. Commun., 25, 2467. (d) Yamazaki, N., Higashi, F., and Iguchi, T. (1974) Tetrahedron Lett., 15, 1191. (e) Yamazaki, N., Iguchi, T., and Higashi, F. (1975) Tetrahedron, 31, 3031. (f) Yamazaki, N., Tomioka, T., and Higashi, F. (1976) Bull. Chem. Soc. Jpn, 49, 3104. (g) Yamazaki, N., and Yamaguchi, M. (1979) Synthesis, 355. (h) O’Brien, R.A., Worman, J.J., and Olson, E.S. (1992) Synth. Commun., 22, 823.

References (i) Uhrich, K., Olson, E.S., and Worman, J.J. (1986) Synth. Commun., 16, 1387. (j) Porwanski, S., Menuel, S., Marsura, X., and Marsura, A. (2004) Tetrahedron Lett., 45, 5027. (k) Chaturvedi, D., Mishra, N., and Mishra, V. (2008) Monatsh. Chem., 139, 267. 127 (a) Fichter, F. and Becker, B. (1911) Ber. Dtsch. Chem. Ges., 44, 473. (b) Morino, S., Sakai, M., Kashiki, I., Suzuki, A., and Miki, M. (1978) Hokkaido Daigaku Suisangakubu Kenkyu Iho, 29, 75. (c) Mulvaney, J.F. and Evans, R.L. (1948) J. Ind. Eng. Chem., 40, 393. 128 (a) Shi, F., Deng, Y., SiMa, T., Peng, J., Gu, Y., and Qiao, B. (2003) Angew. Chem., Int. Ed., 42, 3257. (b) Jiang, T., Ma, X., Zhou, Y., Liang, S., Zhang, J., and Han, B. (2008) Green Chem., 10, 465. 129 (a) Morimoto, Y., Fujiwara, Y., Taniguchi, H., Hori, Y., and Nagano, Y. (1986) Tetrahedron Lett., 27, 1809.

130

131

132

133

134

(b) Tai, C.C., Huck, M.J., McKoon, E.P., Woo, T., and Jessop, P.G. (2002) J. Org. Chem., 67, 9070. (a) Nomura, R., Yamamoto, M., and Matsuda, H. (1987) Ind. Eng. Chem. Res., 26, 1056. (b) Nomura, R., Hasegawa, Y., Toyosaki, T., and Matsuda, H. (1992) Chem. Express, 7, 569. (c) Nomura, R., Hasegawa, Y., Ishimoto, M., Toyosaki, T., and Matsuda, H. (1992) J. Org. Chem., 57, 7339. (a) Sasaki, Y. and Kagawa, H. (1989) Kogai Shigen Kenkyusho Iho, 18, 13. (b) Fournier, J., Bruneau, C., Dixneuf, P.H., and Lécolier, S. (1991) J. Org. Chem., 56, 4456. Shi, F., Zhang, Q., Na, Y., He, Y., and Deng, Y. (2005) J. Am. Chem. Soc., 127, 4182. Seki, T., Kokubo, Y., Ichikawa, S., Suzuki, T., Kayaki, Y., and Ikariya, T. (2009) Chem. Commun., 349. Bazarov, A.I. (1870) J. Prakt. Chem., 2, 283.

167

169

7 Synthesis of Linear and Cyclic Carbonates Danielle Ballivet-Tkatchenko and Angela Dibenedetto

7.1 Introduction

The aim of this chapter is to present the state of the art of the industrial production of major acyclic and cyclic carbonates (Figure 7.1a and b, respectively), and to introduce recent developments in areas of research targeting carbon dioxide (CO2)-based routes. The success of these syntheses has long been based on the reaction between an alcohol or phenol or glycol and phosgene. In 1970, SNPE exploited the phosgenation of glycols that had long been a prominent technology but which, unfortunately, led to the production of chlorinated waste that caused a major adverse environmental impact. Whilst the major positive aspect of using phosgene is its high reactivity, the drawbacks include limitations to transport and storage, safety measures in handling, and the disposal of large amounts of end-products [1]. Environmental regulations and risk minimization, however, do not ensure that such technology can ﬁt in with the market expansion of the two major acyclic carbonates, namely dimethyl carbonate (DMC) and diphenyl carbonate (DPC), and that of ethene carbonate (EC) and propene carbonates (PC). In reality, the industrialization of alternative routes has been developed mainly by EniChem, Ube Industries, and Asahi Kasei, all of which have patented processes based on the use of carbon monoxide. Yet, this favorable situation can be foreseen for other carbonates. Today, less- toxic starting materials such as CO2 are under investigation, mainly on the basis that to shift from using phosgene to using CO2 means that a much safer, but less reactive species (for COCl2 ΔGf0 = −204.9 kJ mol −1, while for CO2 ΔGf0 = −394.4 kJ mol −1 ), is involved. This in turn means that the use of CO2 will require the development of an “ad hoc” catalyst capable of activating the process. In this chapter the processes used for the synthesis of either acyclic or cyclic carbonates, as alternatives to the use of phosgene, will be considered. Details of both the rich patent and scientiﬁc literature that is appropriate to the subject will also be included.

Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

170

7 Synthesis of Linear and Cyclic Carbonates

R

O R

O

O

R'

O O (b)

(a) Figure 7.1

O

(a) Acyclic and (b) cyclic organic carbonates. R, R′ = alkyl, aryl.

7.2 Acyclic Organic Carbonates 7.2.1 Market and Production

Acyclic organic carbonates ﬁnd their major applications as intermediates for pharmaceuticals, agrochemicals, and engineered polymers [2, 3]. However, the market potential is far from having been fully exploited. As end-products, acyclic organic carbonates exhibit valuable properties as lubricants [4], electrolytes for lithium-ion batteries [5, 6], solvents for coating [7], varnish [8] and catalytic reactions [9–11]. In the energy sector, fuels blended by acyclic carbonates lead to better combustion and reductions in emissions [12–14]. Since, for such applications, production needs worldwide are anticipated in the range of several tens of megatons each year, a major scale-up of acyclic carbonates production will become necessary. Yet, it is doubtful that the current on-stream technologies will satisfy this demand. Today, although the phosgene route has a leading role in the production of a variety of organic carbonates, obvious technical and economical barriers continue to be raised with regards to its increased availability and use. Phosgene (COCl2) is widely recognized as one of the most acutely toxic substances used in commerce today. It is currently produced from brine and CO in a two-step process: (i) an electrolysis of brine to produce Cl2 (the chlorine-alkali processes); and (ii) a catalytic reaction with CO on activated charcoal catalysts to forming COCl2 [15]. Notably, among the typical applications of organic carbonate synthesis, the reactions lead to the coproduct HCl; thus, the chlorine cycle Cl− → Cl2 → Cl− is, in overall terms, energy-intensive (Scheme 7.1). Today, phosgene-free technologies to DMC are already entering commercial application, increasing the production volume from specialty applications to larger-scale use as an intermediate in polycarbonate processes. In 2005, the worldwide consumption of DMC was about 100 000 tons, with production concentrated mainly in the United States, Europe, and Japan. Yet, today, DMC production is expanding rapidly in other countries, with China forecasting a production capacity in the range of 20 000 tons [16]. Currently, much attention is being focused on DMC applications due to its low toxicity, an absence of any irritant or mutagenic effects, its high biodegradability [17], and low atmospheric loss [18]. Consequently, DMC is today considered as a chemical feedstock that fulﬁls the “green” chemistry

7.2 Acyclic Organic Carbonates

brine electrolysis

Cl2

CO

COCl2

2 CH3OH

H3OC

C

OCH3 + 2 HCl

O Scheme 7.1 The chlorine cycle for DMC synthesis via the phosgene route.

and engineering criteria [19, 20] that are aimed at handling safer chemicals, and at selecting the safest and most efﬁcient way in which to synthesize those chemicals. With one carbonyl group and two methoxy groups, DMC may be used for carbonylation and methylation procedures, replacing previously used high-risk and environment-damaging compounds such as phosgene, dimethyl sulfate, and halomethane [17, 21–23]. The three commercial routes to DMC are based on the following chemical equations: 2 CH3OH + COCl2 → DMC + 2 HCl

(7.1)

2 CH3OH + CO + 1 2 O2 → DMC + H2O

(7.2)

2 CH3OH + (CH2O)2 CO → DMC + HO (CH2 )2 OH

(7.3)

Nowadays, two of these routes are based on phosgene-free catalytic technologies. The oxidative carbonylation of methanol (Equation 7.2) was introduced on an industrial scale during the early 1980s by EniChem [24, 25]; thereafter, during the early 1990s, Ube Industries Ltd (Japan) developed a variation on the basis of a two-stage process [26]. The ﬁnal catalytic technology, which was introduced by Asahi Kasei in the early 2000s, consists of a transesteriﬁcation between methanol and ethylene carbonate (Equation 7.3) [27]. Interestingly, these examples show that every ten years, a technological breakthrough is witnessed from innovative chemical and engineering concepts. The EniChem DMC process is based on the liquidphase oxidative carbonylation of methanol in the presence of copper chloride as catalyst. The reactor runs under reasonable conditions (373–403 K, 2–3 MPa), with selectivity towards DMC in excess of 95%, and the main byproducts being methyl chloride, dimethyl ether, and CO2. The conversion per pass is limited, partly due to catalyst deactivation by the coproduced water. Additional constraints include the presence of chloride, which leads to a need for corrosion-resistant reactors, as well as the nature of reactants, CO and O2, which implies a need for good kinetic control. In principle, the use of solid catalysts should overcome the corrosion problems and improve product recovery. The gas-phase reactor technology implemented by Ube Industries utilizes an heterogeneous catalyst based on PdCl2, supported on activated carbon. Here, besides the change in catalyst nature, the main modiﬁcation resides in a two-stage continuous process in which methanol is ﬁrst reacted with NO at ∼323 K to produce the intermediate methyl nitrite and water. In the second stage, methyl nitrite is converted to DMC and NO by the addition of CO under catalytic conditions (383–423 K, 0.1–2 MPa). The NO is further recycled into the ﬁrst reactor. The selectivity for DMC lies in the range

171

172

7 Synthesis of Linear and Cyclic Carbonates

90–95%, based on CO and MeONO consumptions. Notably, the process not only eliminates the handling of slurries required by the EniChem process, but also prevents catalyst deactivation by water. Unfortunately, however, both processes require the manipulation of toxic and hazardous reactants that hampers the large scale-up of this technology. By comparison, the catalyzed transesteriﬁcation reaction between ethylene carbonate and methanol (Equation 7.3) offers an alternative for “greening” DMC production. In this Asahi Kasei process [27], the preferred catalyst is based on an anion-exchange resin operating under catalytic distillation conditions between 333–353 K. This reactor design shifts the thermodynamic equilibrium towards complete conversion of ethylene carbonate, such that both the yield and selectivity for DMC and monoethylene glycol are 99.5%. The process is capable of supplying monoethylene glycol to the market, and DMC for captive use to produce DPC. Diphenyl carbonate is an important intermediate in the production of bisphenolA-polycarbonate (BPA-PC). The technology shift from phosgene to DPC to produce BPA-PC on a commercial scale also allows the number of applications of this type of polymer to be increased; for example, high-performance BPA-PCs utilized for information storage (e.g., DVDs) are prepared from high-purity DPC. The chemical route to this DPC brand is based on a two-step reaction, namely transesteriﬁcation followed by disproportionation (Equations 7.4 and 7.5). C6H5OH + (CH3O)2 CO → (CH3O)(C6H5O) CO + CH3OH

(7.4)

2 (CH3O)(C6H5O) CO → (C6H5O)2 CO + (CH3O)2 CO

(7.5)

From a thermodynamics basis, the transesteriﬁcation reaction favors the formation of methylphenyl carbonate (Equation 7.4), whilst its further disproportionation in a second-stage continuous reactive distillation column affords DPC with selectivity >99%. Although both reactions occur at a relatively high temperature (∼ 473 K), optimization of the reaction conditions and engineering design would allow a productivity that ﬁtted with the economics [17, 27]. The multistep integrated Asahi Kasei technology from EC to BPA-PC, which was industrialized during the early 2000s, is a modern, successful example of the exploitation of CO2 within the chemical industry, with EC being produced from CO2 and ethylene oxide. Moreover, it brings “green chemistry and engineering” into practice [28]. The key role played by DMC and CO2 in the practice of green chemistry has led to a stimulation of research for developing other innovative catalytic methodologies for the synthesis of acyclic carbonates, and these are summarized and discussed in the following sections. 7.2.2 Current Trends with CO2 as Feedstock

Conceptually, two routes are available to acyclic carbonates from CO2, the so-called “indirect” and “direct” routes. The ﬁrst strategy is to operate in stepwise fashion,

7.2 Acyclic Organic Carbonates

taking advantage of commercial feedstocks made from CO2 such as ethylene (or propylene) carbonate and urea. Alcoholysis with a monoalcohol, ROH, can then be applied to supply the acyclic carbonate (Equations 7.6 and 7.7). 2 ROH + cyclic carbonate → (RO)2 CO + glycol

(7.6)

2 ROH + urea → (RO)2 CO + 2 NH3

(7.7)

These stepwise reactions raise yield and selectivity issues in the ﬁnal acyclic carbonates, due to thermodynamics constraints. The second strategy, the so-called “direct” route, consists of converting CO2 in one-step to create an end-product (Equation 7.8). This reaction is also thermodynamically equilibrated (see below). 2 ROH + CO2 → (RO)2 CO + H2O

(7.8)

A comparison of the different chemical equations to acyclic carbonates (Equations 7.1–7.8) highlights the fact that, as different coproducts are formed (HCl, glycol, ammonia or water), it might be interesting to assess the efﬁcient use of the reagents according to Green Chemistry principles [19]. The atom economy (AE) factor (expressed in wt%) is the most helpful indicator for that purpose [29], and the AE values for obtaining DMC from the ﬁve stoichiometric reactions mentioned above are listed in Table 7.1. The phosgene route (entry 1) gives AE = 55 wt%; this indicates that, with every ton of DMC produced, 0.8 ton of HCl is also created. A better AE value (59 wt%) is obtained with the coproduction of ethylene glycol from the methanolysis of ethylene carbonate (entry 2). The AE value is higher when water is coproduced (83 wt %, entries 4–5). The use of urea as a feedstock (entry 3) may also lead to AE = 83 wt%, by coupling the urea synthesis equation. The complete use of feedstock atoms (AE = 100 wt%) has been demonstrated commercially for cyclic carbonate synthesis, but as yet no real case has been reported for unstrained ethers (entry 6). The intense patent activity in this area attests to the vast interest for implementing nonphosgene technologies to produce organic carbonates [12]. Hence, the state of the art in research for the three chemical routes based on CO2 (i.e., transesteriﬁcation, urea alcoholysis, and direct carbonation) is discussed in the following sections.

Table 7.1

Atom economy of different chemical routes to DMC.

Entry

Feedstock

Product

AE (wt%)

1 2 3 4 5 6

COCl2 + 2 CH3OH (CH2O)2CO + 2 CH3OH (NH3)2CO + 2 CH3OH CO + 1/2O2 + 2 CH3OH CO2 + 2 CH3OH CO2 + CH3OCH3

DMC + 2 HCl DMC + HO(CH2)2OH DMC + 2 NH3 DMC + H2O DMC + H2O DMC

55 59 72 83 83 100

173

174

7 Synthesis of Linear and Cyclic Carbonates

7.2.3 Alcoholysis of Urea

The alcoholysis of urea affords stepwise acyclic carbonates through the formation of an alkyl carbamate intermediate (Scheme 7.2). The reaction of urea with alcohols to carbamate is exothermic, whereas the subsequent reaction – carbamate to carbonate – is endothermic. Thus, the ideal gas free energy, ΔG, is positive for the latter step, which means that a low yield of carbonates would be expected [12]. Hence, a continuous elimination of the NH3 thus formed would contribute to a favorable shift of the equilibrium to carbonates in the presence of catalysts, so as to enhance the reaction rate. Yet, parallel reactions do occur that lead to side products; for example, decomposition of the carbamate may lead to isocyanic acid and ammonia, while carbonate produces ether and CO2. Moreover, as the acyclic carbonate formed is an excellent alkylating agent, then the N-alkylation of urea and alkyl carbamate would also occur if acyclic carbonate were to accumulate in the reaction. Consequently, reactor engineering and catalyst design are of primary importance to circumvent these secondary reactions and to eliminate the product dialkyl carbonate from the reaction mixture on continuous basis. For such a purpose, a catalytic distillation set-up [30, 31] that combines the chemical synthesis with separation by distillation represents a well-suited technology. Accordingly, the process intensiﬁcation has been effective, offering the advantage of moving in the direction of “green” engineering principles [20]. A very recent report on the modeling of catalytic distillation processes for the synthesis of DMC via urea methanolysis illustrates this technology [32]. Whilst the concept of catalytic distillation is simple, its practice is complex and implies a need for process modeling, for monitoring the thermodynamics of the ﬂuid-phase equilibria, and for optimizing the chemical kinetics. Catalyst design, therefore, plays a major role, and the use of either soluble organotin compounds [33–35] or oxides, for example, ZnO [36], CaO, MgO, ZrO2 [37], have been reported. Side reactions may also produce primarily N-alkylated products. Among soluble precursors, dibutyltin(IV) provides the most active system, with the yield and selectivity being drastically enhanced by running the reaction in a reactive distillation reactor with triethylene glycol dimethyl ether as a high-boiling solvent. The DMC is eliminated continuously from the boiler (453 K, 1–1.5 MPa), with selectivity and conversion in excess of 98% [38]. The details of an exhaustive kinetic study have been reported [35]. The preparation of higher acyclic carbonates is also feasible from urea alcoholysis with primary and secondary alcohols [39–43], with high reaction temperatures of up to 543 K having been reported.

H2N

C

NH2

O Scheme 7.2

ROH -NH3

H2N

C O

OR

ROH -NH3

RO

C

OR

O

Stepwise reaction to acyclic carbonates from urea and alcohol.

7.2 Acyclic Organic Carbonates

CH3OH

Bu Bu

Sn

OCH3 OCH3

H2NC(O)OCH3 CH3OH

Bu Bu

Sn

OCH3

(CH3O)2CO

Bu

N OCH3 H C

Bu

Sn

OCH3 NH2

O

NH3

H2NC(O)OCH3

Scheme 7.3 Key reactions proposed for the formation of DMC from methyl carbamate and methanol with n-Bu2Sn(OCH3)2 precursor.

To the best of the present authors’ knowledge, correlations between the catalyst’s structure and activity have been demonstrated only for tin catalysts [41]. A series of di-n-butyl tin(IV) compounds have been synthesized, characterized by nuclear magnetic resonance (NMR) and infrared (IR) spectroscopies, and screened for methyl carbamate methanolysis at 463 K. The key reactions proposed are depicted in Scheme 7.3. Di-n-butyldimethoxy stannane precursor reacts with methyl carbamate to afford dibutylmethoxymethylcarbamato stannane, n-Bu2Sn(OCH3)[HNC(O)OCH3], for entering the catalytic cycle. A nucleophilic attack of the alcohol on the carbon atom of the C=O fragment then takes place for DMC elimination, with concomitant formation of the dibutylmethoxyamino tin(IV) intermediate. This intermediate quickly reacts with methyl carbamate, releasing ammonia and regenerating n-Bu2Sn(OCH3)[HNC(O)OCH3]. A stimulating development of urea alcoholysis has been demonstrated very recently for better AE, in an innovative integrated process that incorporates fatty ester hydrolysis to ω-amino-alkanoic acids [44]. Within the scope of this chapter, the most interesting step of this process is the recycling of waste alcohol, formed by the hydrolysis step, for urea alcoholysis. Dialkyl carbonate is produced together with ammonia; thereafter, the ammonia is engaged in the amination reaction to obtain the amino acids. The overall process avoids the storage of NH3 that is necessary for the amination route, and transforms a waste product – the alcohol – into the valuable dialkyl carbonate. 7.2.4 Direct Route

The alkylation reaction of various alkali and alkaline-earth metal carbonates with alkyl halides R(CH2)nX (X = Cl, Br, I) is a primary synthetic procedure in organic chemistry for obtaining various symmetrical and unsymmetrical dialkyl carbonates under phase-transfer conditions in polar aprotic solvents [45]. Excellent yields may be obtained by running the reaction at 383 K in ionic liquids such as

175

176

7 Synthesis of Linear and Cyclic Carbonates

imidazolium salts, offering the advantage of simple experimental and work-up procedures [46]. An alternative route from CO2, alcohols and alkyl halides is also effective by the addition of a strong base (e.g., substituted guanidines), so as to trap the hydrogen chloride that is formed [47]. In this reaction (which must be conducted under pressure), the alcohol is deprotonated by the base (Equation 7.9), which leads to a hemicarbonate species ROC(O)O− (Equation 7.10) that is then further alkylated by the alkyl halide to give the acyclic carbonate (Equation 7.11). ROH + B ( base) → RO-BH+

(7.9)

RO-BH+ + CO2 → ROC (O) O-BH+

(7.10)

+

ROC (O) O-BH + RX → (RO)2 CO + BHX

(7.11)

RO-BH+ + RX → ROR + BHX

(7.12)

In order to promote and stabilize the hemicarbonate anion versus the alkoxide (Equations 7.9 and 7.10), the reaction is generally run under CO2 pressure so as to improve the carbonate yield compared to the corresponding ether (Equations 7.11 and 7.12). The reaction is not always effective, however, with attempts to prepare diaryl carbonates having led exclusively to diaryl ethers. This methodology has also been applied to DMC synthesis from CO2, methanol and methyl iodide in the presence of K2CO3. In this case, both DMC and dimethyl ether (DME) are produced [48], while K2CO3 is transformed into KI during the course of the reaction, which makes this synthetic protocol barely catalytic [49, 50]. In the absence of alkyl halides, the reaction of alcohols and CO2 to dialkyl carbonates is quantitative when a stoichiometric amount of diethylazodicarboxylate/PPh3 [51] or dicyclohexylcarbodiimide (DCC) is added [49, 52]. One of the very ﬁrst reports of a catalytic reaction (Equation 7.8) involved the use of di-n-butyldialkoxy stannanes, n-Bu2Sn(OR)2 (R = methyl, ethyl n-butyl) as catalytic precursors [53]. The best yield obtained was for diethyl carbonate (DEC) from ethanol under 1 MPa of CO2 in a batch reactor at 443 K for 24 h. The DEC : Sn molar ratio thus obtained was equal to 6.6. Later, this reaction was extended to DMC synthesis with the same stannanes at 423 K under 2.8 MPa pressure, when the DMC : Sn molar ratio was found to increase to 3 in the presence of DCC used as a chemical scavenger of the water that had been co-produced [54]. Acetals (e.g., 2,2-dimethoxypropane, DMP) are also good candidates for chemical water trapping, invariably leading to an increase in the turnover numbers (TONs) [55–57]. Interestingly, the coaddition of triﬂate salts to n-Bu2SnO (Sn : triﬂate = 10) caused a signiﬁcant improvement in the TON, from 4 up to 10, at 453 K under 30 MPa for 24 h [58]. One drawback of chemical trapping is that either a substituted urea (from DCC) or acetone (from DMP) are coproduced. Physical water trapping represents a cleaner alternative that has been demonstrated with molecular sieves 3 or 4 Å [59–61]. In order to produce more active and recyclable tin-based catalysts, both polystyrene-grafted organotin species [62] and immobilization on mesoporous SBA-15 [63] have been employed. Although recyclable, these systems require

7.2 Acyclic Organic Carbonates

optimization for better activity, and much attention has been paid over the past decades to utilizing soluble n-Bu2Sn(IV) derivatives as catalysts for the formation of DMC from CO2 and methanol, in either the patent or the open literature. The most likely reasons for this are the 100% selectivity to DMC, and the robust nature of the n-Bu2Sn(IV) moiety under the reaction conditions. Nonetheless, other soluble alkoxides such as those of titanium(IV) [64, 65] and Group V metals [66, 67] all exhibit activity for DMC formation. Heterogeneous catalysts also prompt the formation of DMC from methanol and CO2. Among these, ZrO2-based materials have been thoroughly studied, with the reaction temperature lying in the range 403–443 K, similar to the soluble complexes. The selectivity to DMC was close to 100% [68], and the modiﬁcation of ZrO2 with H3PO4 enhanced the rate of DMC formation [69]. Even more active was the Ce0.2Zr0.8O2 mixed oxide, which showed no detrimental effects on selectivity [70]. In line with this, the activity of pure or doped CeO2 has also been reported [71, 72]. As noted with organotin catalysts, the addition of a chemical water trap (e.g., DMP) can increase the conversion drastically [73]. However, DME formation did occur at a high DMP loading, most likely due to DMP decomposition. Increasing the reaction pressure to supercritical conditions has a beneﬁcial effect on alumina and titania catalysts modiﬁed by sulfate or phosphate ions [74]. Efforts to design catalysts that are more efﬁcient at low temperature, so as to take advantage of a more favorable thermodynamic equilibrium, led to the ZrO2and CexTi1−xO2-supported polyoxometallates. The system H3PW12O40/ZrO2 is active at 373 K, but less selective, such that traces of DME and CO are detected [75]. Among the H3PW12O40/CexTi1−xO2 materials tested, H3PW12O40/Ce0.1Ti0.9O2 provided the highest DMC yield, at 443 K, which correlated with the higher number of acid and basic sites. Interestingly enough, unsupported H3PMo12O40 and its copper salt Cu1.5PMo12O40 were active at temperatures as low as 313–333 K under an atmospheric pressure of CO2 [76]. Unfortunately, the DMC selectivity did not exceed 23%, and DME, formaldehyde, and methyl formate were the other products analyzed. A much higher selectivity, close to 85%, was also achieved at atmospheric pressure under UV irradiation of a Cu-doped Ni–V mixed oxide supported on silica [77]. At 403 K, the DMC yield was almost twice that achieved without irradiation. The collection of catalysts developed for the direct carbonation of alcohols to dialkylcarbonates has been enlarged over the past few years, mainly targeting rate enhancement, and a high selectivity (∼100%) has already been achieved with a variety of catalytic compositions. The reaction is generally at its best under high pressure, typically 10–30 MPa, as was recognized long ago [49, 55, 57, 74, 78], the consequence being that the ﬂuid phase equilibrium must be scrutinized. This type of information is essential, as variations in the pressure, temperature and molar fraction can induce changes in the number, composition, and volume of the ﬂuid phases (i.e., liquid, gas, supercritical). One outcome of this has been the inﬂuence on chemical kinetics, such that experimental and predicted ﬂuid phase equilibria for binary, ternary, and quaternary CH3OH/CO2/DMC/H2O mixtures are presently available over the temperature and pressure range of the catalytic reaction

177

178

7 Synthesis of Linear and Cyclic Carbonates

[79–82]. The data collected data also permit simulations to be prepared of the separation steps of the outlet feed, in order to optimize the integrated process reaction– separation. The membrane technology for DMC separation has also been investigated very recently [83], whilst chemical engineering – and, in particular, reactor design – being considered of equal importance to the chemistry of the process if the thermodynamic equilibrium is to be shifted favorably, as can be achieved with a membrane catalytic reactor [84]. The enthalpy and free energy calculations of the reaction from methanol (Equation 7.8) show that the reaction is slightly exothermic, but does not occur spontaneously due to ΔG > 0. Whilst the available data are obtained at a constant temperature and pressure, and are for ideal ﬂuids [12, 85, 86], this clearly this does not correspond to the “real” situation. Indeed, in one study it was shown that tuning the phase behavior had a signiﬁcant impact on the equilibrium for DMC synthesis [87]. Besides thermodynamics, the low conversion to dialkyl carbonates may stem from the poisoning of the catalyst by water. The recycling of catalysts has also been demonstrated [62, 80, 88], and this may offer additional technological solutions, as evidenced by very recent patents ﬁled by the Asahi Kasei company [89–92]. The reaction mechanism behind the direct carbonation of alcohols has been investigated not only over ZrO2, but also with the soluble complexes Nb(OCH3)5 and R2Sn(OCH3)2 (R = CH3, n-Bu). The main features of the catalytic cycles are described in the following sections. With ZrO2 [93] and H3PO4/ZrO2 heterogeneous catalysts [69], NMR, IR and Raman spectroscopic studies led to the proposal of a reaction sequence that is summarized in Scheme 7.4. Surface Zr–OH groups and Zr4+O2− sites may act as Lewis acid–base pairs for the bifunctional activation of methanol to form water and CH3O–Zr species. A subsequent CO2 insertion then occurs, converting the methoxy fragment to the hemicarbonate CH3OC(O)O–Zr species, after which CH3OH activation on acid sites for methyl transfer to CH3OC(O)O–Zr leads to DMC. It has been suggested that Brønsted acid sites may be more effective than Lewis acid sites, due to better results having been obtained with H3PO4-modiﬁed ZrO2. With soluble catalysts, both experimental investigations and density functional theory (DFT) calculations with a soluble Nb(OCH3)5 catalyst [66, 94] support the

H O Zr

CH3OH O

H 2O

Zr Zr

O

CH3 O Zr

DMC OCH3 CH3OH Zr

O

O

C

CO2 O

Zr

Scheme 7.4 Postulated key surface species involved for DMC formation from CH3OH and CO2 over ZrO2.

7.2 Acyclic Organic Carbonates

CO2 (CH3O)4NbOCH3 (CH3O)4NbOC(O)OCH3 DMC + H2O OCH3 O C O

2 CH3OH

(CH3O)4Nb CH3O H

OCH3 H

Scheme 7.5 Postulated cycle for DMC formation from CH3OH and CO2 with Nb(OCH3)5.

Scheme 7.6 X-radiography-derived structures of isolated di-n-butyltin(IV) compounds.

most probable reaction pathway, which involves CO2 insertion into one Nb–OCH3 bond, followed by the activation of methanol. Interestingly, two molecules of methanol are involved in the formation of DMC from the hemicarbonate moiety. The ﬁrst of these molecules is coordinated to the nobium center and acts as a Lewis acid, while the second methanol molecule binds to the ﬁrst one via hydrogen bonding (Scheme 7.5). This second activation favors the transfer of a methyl group to the oxygen atom of the C=O fragment of the Nb–OC(O)OCH3 moiety. Further intramolecular rearrangement then causes the elimination of DMC and water, and leads back to Nb(OCH3)5. With tin compounds, the solution and solid-state structural characterization of intermediates has been achieved from (CH3)2Sn(OCH3)2 [95] and n-Bu2Sn(OCH3)2 [78, 80, 96, 97], and n-Bu2Sn[OCH(CH3)2]2 [98]. Three isolated compounds are considered as resting species in the catalytic cycle (Scheme 7.6). Here, compounds A and B are obtained from the insertion of CO2 into the Sn–OR bonds of the corresponding methoxy precursors. Under CO2, thermal treatment transforms A into B, with the concomitant formation of DMC, whereas species B gives C and DMC only in the presence of methanol. It is worth noting here that C is the species which can be recycled many times for DMC formation, without any loss of activity.

179

180

7 Synthesis of Linear and Cyclic Carbonates

The transformation of A to B to C underlines an increase in the nuclearity of the resting species with the formation of Sn–O–Sn linkages, the tin centers being pentacoordinated. It is worth noting that monomeric species are most likely involved in the catalytic cycle. A comparison between the proposed key steps for DMC synthesis from methanol and CO2 on ZrO2-based and soluble catalysts highlights their similarities. The CO2 reactivity towards M–OR bonds leads to the hemicarbonate M–OC(O)OR, and for this a bifunctional methanol activation is a prerequisite. It is ﬁrst necessary to produce CH3O methoxy species for providing the M–OCH3 fragments prompted to insert CO2. Capture of the proton of methanol is effective with basic sites of the catalyst. Second, the hemicarbonate M–OC(O)OR is alkylated by a methyl group arising from methanol activation on acid sites. Clearly, the proposed sequence implies an electronic and steric tuning of the network of acid–base sites in order to optimize the rate of the catalytic cycle. 7.2.5 The Future of CO2-Based Routes to Acyclic Carbonates

The expanding market for acyclic carbonates over a wide range of applications is, unambiguously, due to the commercialization of phosgene-free processes. These catalytic processes respond not only to the demand of greener chemistry and engineering, but also to the demand of higher quality products, as exempliﬁed by BPA-PC production. Despite the CO2-based routes suffering from unfavorable thermodynamics, the commercial applications bear witness to favorable economics as a result of breakthroughs in process engineering. The indirect CO2-based transesteriﬁcation route represents the latest method for achieving high-purity DMC and DPC production. In the case of DMC synthesis, the monoethylene glycol that is coproduced ﬁts with the market strategy of avoiding the handling of harmful waste materials. Although DPC is an intermediate in BPA-PC production, whether the DMC-DPC technology will be used solely for the worldwide production of BPA-PC (ca. 2.7 Mt per year) is unclear; nonetheless, the DMC market will surely continue to expand to the mega-tonne scale. The alcoholysis of urea is another indirect CO2-based route to dialkyl carbonates that has been investigated with assistance from catalytic distillation technology. This route has the potential to be integrated with a urea facility for optimizing ammonia recycling, and hence the reaction mass balance would correspond to that for DMC synthesis from methanol and CO2. It worth noting that large-scale DMC production would beneﬁt from urea plant capacities (ca. 120 Mt per year worldwide). The main chemistry issue would be selectivity and catalyst stability, due to the high reaction temperatures required. A better understanding of the reaction mechanism is necessary for catalyst optimization. Although the direct CO2-based route to dialkyl carbonates offers selectivity improvement (almost 100% for DMC synthesis), limited yields may result due to unfavorable thermodynamics and catalyst deactivation by water. Until now, mechanistic studies have been fruitful in identifying some of the key steps of the

7.3 Synthesis of Organic Cyclic Carbonates

ethylene from oil refinery

CO2

O

O

O C

- HO

OH

O H2 + CO/CO2 from fossil fuels or biomass

H2 + N2

CO2 -H2O

2 CH3OH

H3CO

C

OCH3

O

2 NH3

CO2 -H2O

H2 N

C

NH2

- 2 NH3

O

Scheme 7.7 Flow chart of the transesteriﬁcation, urea methanolysis, and direct carbonation for DMC.

catalytic cycle for better catalyst design. Whilst water-trapping experiments unambiguously beneﬁt yield increases, an alternative approach is that of catalyst recycling, which has the advantage of avoiding the cogeneration of chemicals during chemical water trapping. Further development in reactor design are needed to assess the potential of the direct carbonation of alcohols for commercial application. The advantage of a direct reaction is that it is independent of the demand and supply balance of the coproduct (glycol or ammonia), which forms part of the indirect route. Feedstock supply is also an issue. A ﬂow chart comparing the three technologies is shown in Scheme 7.7. Methanol is currently produced from syngas, which has been enriched with CO2 and is derived from either a fossil fuel or biomass. A better approach might consist of the direct hydrogenation of pure CO2, whilst transesteriﬁcation with ethylene carbonate implies that ethylene, derived from oil reﬁneries, should be used as the raw material. The urea route does not involve any speciﬁc carbon source for hydrogen and CO2, as does that of methanol synthesis. Ideally, hydrogen should be produced chieﬂy from water splitting, and CO2 from ﬂue gas.

7.3 Synthesis of Organic Cyclic Carbonates

The ﬁrst reports describing the formation of cyclic carbonates (CCs) appeared during the early 1930s [99, 100], whilst the ﬁrst patents (essentially related to the synthesis of ethene and propene carbonates) appeared more than 50 years ago [101, 102]. The Huntsman Corporation is one the world’s largest producers of alkylene carbonates, with a capacity of 33 kt per year, covering approximately 50% of CC production worldwide. Today, CCs are widely used in the manufacture of

181

182

7 Synthesis of Linear and Cyclic Carbonates

products including solvents [103], paint-strippers [104], lithium batteries [105], and biodegradable packaging. They also have applications in the chemical industries [106] and in medicinal chemistry [3]. 7.3.1 Carboxylation of Epoxides

Several reports have been made and patents granted on the formation of CCs from CO2 and epoxides. The latter materials represent the cost drivers in the production of carbonates; that is, propene oxide (PO) used in the synthesis of propene carbonate (PC) is obtained via the oxidation of propene, using hydrogen peroxide. However, a cheaper technology for producing either PO [107–114] or PCs, starting from propene [115, 116], is currently being actively sought. 7.3.1.1 Use of Conventional Solvents The carboxylation of epoxides (Equation 7.13) has long been known (IG Farben, 1943) [12, 117, 118], and today is available on stream from several production plants.

R H

R

H O

H

+ CO 2

cat

H

H O

O C

H (7.13)

O The reaction of epoxides with CO2 affords either CCs or polymers [119], and many reports have been made [120–125] and different active catalysts described [126– 130] such as alkyl ammonium-, phosphonium-salts and alkali metal halides, in this respect. The main drawbacks here are the need for a high catalyst concentration, a high pressure (5 MPa of CO2), and a temperature ranging from 370 to 400 K. The recovery of the catalysts for reuse is also a key issue, and in order to simplify the recovery process various hybrid systems have been developed, an example being that prepared by coupling 3-(triethoxysilyl)propyltriphenylphosphonium bromide with mesoporous silica [131]. In this case, the reaction was carried out in the absence of solvent, under very mild conditions (1 MPa, 263 K, 1 mol% loading of catalyst, 6 h), such that the hybrid catalyst could be recovered and recycled several times. High yields of CCs at atmospheric CO2 pressure, using main group metal halide salts, have also been reported [132]. It has also recently been found that, by using potassium halide as catalyst in the presence of β-cyclodextrin (βCD), CCs can be formed in high yield without the use of an organic solvent [133]. Recently, βCD has also been shown to act as a hydrogen-bonding agent and to accelerate the ring-opening, while the halide served as the active catalyst. Organometallic species and metal complexes [134], classical Lewis acids [122] and metal phthalocyanines [135] have each been used as catalysts. Likewise, heterogeneous catalysts such as

7.3 Synthesis of Organic Cyclic Carbonates

metal oxides [136–140], supported ammonium salts [141, 142] and metal complexes [143] have been used, as these are characterized by a longer life as homogeneous catalysts. Very often, amides such as dimethylformamides (DMF) or dialkylacetamides (DAA) have been used as solvents in reactions where they may themselves promote the carboxylation of epoxides [139], if only to a limited extent. Interestingly, it has been reported [140] that, when starting from pure enantiomers of epoxides, optically active carbonates can be obtained with a total retention of conﬁguration. Conversely, when a racemic mixture of the epoxide was used, an enantiomeric excess (ee) of the order of 22% was obtained using Nb(IV) complexes with optically active (N, O, P as donor atoms) ligands: this low ee-value was due to de-anchoring of the ligand from the metal center, as conﬁrmed using NMR [140]. As noted above, the carboxylation of epoxides may afford polycarbonates, whether using Al–porphyrin complexes [144, 145] or Zn-compounds [146]. In fact, the Al-catalysts, which were the ﬁrst to be described, are currently used in production plants (see Chapter 8). The palladium-catalyzed ﬁxation of CO2 is also a useful method for the synthesis of CCs, with the ﬁrst such example using vinyl-substituted epoxides and a palladium catalyst [Pd(PPh3)4/PPh3] having been reported independently by both Fujinami and Trost [147, 148]. Here, the carbonate is produced via the formation of a ﬁrst π-allyl-palladium intermediate (I); this then ﬁxes CO2 to form a second π-allyl-intermediate (II) that produces vinyl-substituted cyclic carbonates (Scheme 7.8). Functionalized CCs can also be obtained in good yields under mild conditions from epoxides and CO2 by using an electrochemical procedure [149, 150]. For this, the CC formation is catalyzed by Ni(cyclam)Br and is carried out in singlecompartment cells ﬁtted with a magnesium anode. The presence of functional groups such as chlorine, bromine, ether, ester or oleﬁns is compatible with the reaction conditions. O O

O O Pd(0) O-

O

Pd+ O

-

Pd+ O (I)

(II) CO2

Scheme 7.8 Formation of cyclic carbonate using a Pd-catalyst.

183

184

7 Synthesis of Linear and Cyclic Carbonates

7.3.1.2 Use of Ionic Liquids Recently, the carboxylation of epoxides was carried out in ionic liquids (ILs) that demonstrate interesting characteristics such as thermal and chemical stability, selective solubility towards organic and inorganic materials, and a high reusability of the catalysts. Taken together, these parameters make ILs useful for this type of application [151–156]. It is worth mentioning here that CO2 is easily dissolved into the IL phase, which in turn makes the reactions of CO2 possible, and also suitable. In fact, ILs have been reported as one of most efﬁcient media for CO2 ﬁxation in the production of CCs from epoxides [157–162]. In such a case, the catalytic activity may be affected by the presence of water or air. Recently, Sun et al. [163] showed that when hydroxyl groups were added to traditional ILs, both the efﬁciency of the catalyst and the CC yield were increased. By considering water to be an hydroxyl group-containing solvent, Sun et al. [164] have developed a method for preparing CCs in water. For this, PO was used to develop a comparative analysis of the cycloaddition process of CO2 to PO, in the presence and absence of water. In the presence of water, PO was converted into PC with a yield that was four- to ﬁvefold higher than in its absence; moreover, almost all of the Lewis base catalysts used showed a high activity. The cycloaddition of CO2 to other epoxides in water has also been examined using PPh3BuI as catalyst. For this, either aromatic (styrene) or aliphatic (ethene, cyclohexene) epoxides were used, with a high epoxide conversion and ≥92% CC selectivity under the reaction conditions employed (epoxide 0.2 M; H2O 0.067 M; PPh3BuI 1 mM; 2 MPa; 298 K). In the proposed reaction, mechanism three steps have been considered (see Scheme 7.9):

H

H H

O

O H O-

O R

LB+ X

R LBX

O 3

R

CO2

2

1

O

H H O OLB+

R O

X

O O

Scheme 7.9 I−, Cl−.

Reaction path proposed using H2O and PPh3BuI. LBX = Lewis base; X = Br−,

7.3 Synthesis of Organic Cyclic Carbonates

1) An easy opening of the epoxide ring, as water (acidic site) and the bromine anion of the Lewis base (basic site) coordinate different parts of the epoxide. 2) The formation of an alkylcarbonate anion due to an interaction between the oxygen anion and CO2. 3) The formation of a CC via an intramolecular substitution of the halide. 7.3.1.3 Use of Supercritical CO2 Supercritical CO2 (scCO2) is considered to be an economically viable and ecologically benign reaction medium for organic reactions. It has several advantages, such as no ﬂammability, a lack of toxicity, an absence of any gas–liquid phase boundary, and possible simpliﬁcations during work-up. Kawanami et al. have reported that the ﬁxation of CO2 under supercritical conditions effectively proceeds to give CCs (Equation 7.14) [165].

Ph O

ScCO 2-DMF

O

78 atm, 323 K

Ph

(7.14)

O O

For example, styrene oxide has been converted into the corresponding CC in scCO2 with DMF with 85% yield. As reported also by others [139, 140], DMF – when used as a cosolvent – plays a key role in the reaction as it improves the ﬁxation of CO2 into epoxides so as to afford carbonates. This beneﬁcial effect may be due either to a participation in the ring opening of the epoxide, or in a preliminary coordination of CO2. 7.3.1.4 Combined Reaction Media: sc-CO2 and Ionic Liquids A mixed reaction medium, composed of scCO2 and ILs, has been deﬁned as a new biphasic system by Advanced Industrial Science and Technology (AIST), and used for selective and efﬁcient CC synthesis. For example, 1-alkyl-3-methylimidazolium salts represent a suitable system when used under supercritical conditions for the synthesis of CCs [156] from epoxides and CO2. Kanawami et al. [159] have reported that the use of 1-octyl-3-methylimidazolium tetraﬂuoroborate under supercritical conditions resulted in a 100% conversion into PC, with 100% selectivity, within only a few minutes (Equation 7.15).

C8H17 O + ScCO 2 H3C

N

N

CH3

H3C

-

X =BF 4 5 min, 273 K

O

O

100% yield

O (7.15)

185

186

7 Synthesis of Linear and Cyclic Carbonates

Supported ILs also represent a promising alternative in heterogenized catalysis [151, 166–172]. In this case, Jin-Quan Wang et al. [173] supported the IL used by Kawanami [159] onto an amorphous silica (namely, [C4-mim]+X−/SiO2) and found that the solventless (the use of scCO2 avoids the use of other solvents, as it may serve as both reagent and solvent) synthesis of CCs occurred with high yields (78–98%, depending on the epoxide used) and selectivity (78–100%) at a temperature of 433 K under 8 MPa of CO2 and in 4 h. When using supported ILs, the recovery and reuse of the catalyst was simpliﬁed, with only a slight loss in activity being observed after four cycles. 7.3.2 Oxidative Carboxylation of Oleﬁns 7.3.2.1 Use of Oxygen as Oxidant The oxidative carboxylation of oleﬁns appears to be a very interesting synthetic methodology for synthesizing CCs, starting from cheap and easily available reagents such as CO2 and O2 (Equation 7.16).

O2, CO2 RHC=CH2

cat

R H

H O

O C

H

(7.16)

O

The direct oxidative carboxylation of oleﬁns has great potential, and many advantages. Notably, it does not require the CO2 to be free of dioxygen; this is an especially attractive feature, as the cost to purify CO2 is extremely high, and may discourage its use. Moreover, the direct oxidative carboxylation of oleﬁns can couple two processes – the epoxidation of oleﬁns, and the carbonation of epoxides. Hence, the process makes direct use of those oleﬁns that are available commercially at low price, and which represent an abundant feedstock. Such an approach also avoids having to isolate the epoxide. Very few examples have been reported of the direct carbonation of oleﬁns; examples include the direct functionalization of propene [174, 175] and styrene [176, 177]. When using RhClP3 as a catalyst, under homogeneous conditions, Aresta et al. demonstrated [176, 178, 179] the formation of two classes of compounds due to two alternative modes of oxygen transfer to the oleﬁn:

•

One-oxygen addition to the oleﬁn with formation of epoxide and its isomerization products and carbonate (Scheme 7.10, upper part).

•

Two-oxygen addition to the oleﬁn with formation of aldehydes, as an effect of the addition of oxygen to the C–C double bond with cleavage of the double bond of the oleﬁn, and the relevant acids (Scheme 7.10, lower part).

7.3 Synthesis of Organic Cyclic Carbonates

PhCH2CHO PhC(O)CH3 PhHC

CH2

“One oxygen”

H H

to the olefins

O PhHC=CH2

+ CO2

+ O2

"Rh"

transfer

Ph C

H

C

O

Rh = RhClL3

O C

L = PEt2Ph or PEtPh2

O PhCHO

“Two oxygen” transfer

PhCOOH

to the olefins: cleavage of the double bond

Scheme 7.10 Oxidative carboxylation of styrene under homogeneous conditions.

The reaction mechanism has been shown [180] to consist of: (i) interaction of the Rh-catalyst with O2 to afford a dioxygen species; (ii) conversion of the Rh–O2 complex into a peroxocarbonate (Equation 7.17a) by reaction with CO2; and (iii) a one-oxygen transfer to the oleﬁn with formation of the Rh-carbonate which, in principle, should no longer be active as a catalyst (Equation 7.17b). P3RhCl + O2 + CO2 → P3RhCl (CO4 )

(7.17a)

P3RhCl (CO4 ) + L → P3RhCl (CO3 ) + LO

(7.17b)

P3RhCl (CO3 ) + P′ → P3RhCl + CO2 + P′=O

(7.17c)

In fact, P3RhCl has been shown to produce more than 3 mol of carbonate per Rh, thus demonstrating that the Rh(CO3) species can indeed act as a catalyst. Moreover, an oxophile such as a phosphane ligand can extract an O-atom from the coordinated carbonate (Equation 7.17c), thus regenerating Rh(I) and CO2. If the Rh-catalyst is stabilized by a monodentate phosphane ligand, it can be destroyed because the lifetime of the oxidation of phosphine to phosphine oxide is less than 1 h. Yet, if the catalyst is stabilized by a bidentate ligand, and the catalyst has a lifetime of 2–3 h, then it can convert into different species that have no catalytic activity. However, such a synthetic approach would suffer as a result of a short lifetime and a low TON. Several heterogeneous systems (transition metal oxides or oxides from Group I and II elements) [181] have been used as catalysts, with the best results having been obtained when using Nb2O5 [177]. The use of heterogeneous conditions has also shown that oxidation of the oleﬁn does not follow the peroxocarbonate

187

188

7 Synthesis of Linear and Cyclic Carbonates

pathway. More likely, it is a radical process that can be started by the catalyst, which plays a very important role in the carbonation step as the carbonate yield depends on the catalyst used. The selectivity of the process (which reaches a maximum of 50% with respect to the oleﬁn) will still be affected by the formation of byproducts such as benzaldehyde, benzoic acid, acetophenone, phenylacetaldehyde, 1,2ethanediol-1-phenyl, and a benzoic acid ester. Following a short induction time, benzaldehyde is formed in greater amounts than the epoxide, and becomes the predominant product after 45 min. Carbonate formation begins after 1 h and steadily increases with time, while the concentrations of the epoxide and benzaldehyde reach steady status. The catalyst’s life will be in the region of days, and it can be easily recovered when the catalytic run is complete [177]. 7.3.2.2 Use of Other Oxidants Other research groups have studied this synthetic approach using different oxidants rather than O2. For example, Sun et al. [182, 183] reported the one-pot synthesis of styrene carbonate from styrene using tert-butyl hydroperoxide (TBHP) as oxidant, in the presence of quaternary ammonium halides (notably Bu4NBr) or imidazolium salts. The styrene carbonate yield was 38%, although as homogeneous conditions were used this approach presented drawbacks such as a long reaction time and separation and recycling of the catalyst. When Au/SiO2/ ZnBr2 was added as cocatalyst, the styrene carbonate yield was increased to 43% [184]. Eghbali et al. [185] have reported the details of a highly efﬁcient method for converting alkenes and CO2 into CCs directly in water, by using N-bromosuccinimide (NBS) together with 1,8-diazabicyclo[5.4.0]undecen-7-ene (DBU), or a catalytic quantity of bromide ion together with aqueous H2O2. Jing-Lun Wang et al. [186] also used a catalytic system composed of sodium phosphotungstate and n-Bu4NBr (TBAB) to synthesize styrene carbonate in a single step from styrene and CO2, using 30% H2O2 as oxidant. Here, the presence of NaHCO3 was found to improve the formation of styrene carbonate, which was isolated (yield 68%) after a 12 h reaction at 223 K in the presence of 4 equiv. of H2O2, which made the process much less appealing. The synthesis of cyclic carbonates, starting from oleﬁns, can be also carried out via a multistep method based on two separate reactions. To this end, CO2 and the carboxylation catalyst have been added to the same reactor in which a preliminary epoxidation process had been carried out. An example of the synthesis of CCs from oleﬁns in a single reactor has been reported by Srivastava [187, 188], by using titanosilicalite as catalyst and hydroperoxide as oxidant in the form of H2O2 or TBHP. The reaction was carried out in two steps, in which the oleﬁn was ﬁrst epoxidated at 233 K using H2O2 or TBHP. The CO2 was then added in presence of N,N-dimethylamionopyridine as cocatalyst, to afford a 33% yield of the CC at 293 K. Similarly, Ono et al. [189] have reported using a composed catalytic system, namely MTO/UHP/Zn[EMIm]2 Br4/[BMIm]BF4 (UHP, urea hydrogen peroxide and MTO, methyltrioxorhenium). With the multistep method described above, a

7.3 Synthesis of Organic Cyclic Carbonates

189

yield of styrene carbonate of 83% was achieved. It was suggested that, as the use of O2 was more cost-effective and afforded similar, this reaction warranted further investigation and improvement. 7.3.3 Other Synthetic Routes to Cyclic Carbonates 7.3.3.1 From Halohydrins The ﬁrst examples of the synthesis of CCs from halohydrins appeared during the early 1930s. In this case, vic-halohydrins were reacted with sodium hydrogen carbonate [190, 191] or sodium alkyl carbonates [192] to afford CCs in good yield (82%), especially when the reaction was carried out under high pressure and temperature. When tetramethylammonium hydrogen carbonate was used in acetonitrile (Equation 7.18) [193], the relevant carbonate could be prepared in high yields under mild conditions.

R2

R2 H R3 + (CH3)4N+HCO3-

R1 HO

X

R1, R2, R3 = H, CH3, C6H5; X = Cl, Br

CO2/acetonitrile 293 K, 10-75 min

H R3

R1 O

O

+ (CH3)4N+X- + H2O

O

(7.18) − If the halohydrins were added to solutions containing C2O6−2 or CO2− 3 , or HCO3 anions in the presence of tetra-alkylammonium cations, then CCs would have been obtained in good to high yields. Thus, the formation of CCs was seen to require the presence of a suitable leaving group in an α position with respect to the alcoholic hydroxy group (e.g., halohydrins) [194]. It has also been shown that epihalohydrin may react with K2CO3 to afford the corresponding ﬁve-membered CCs containing an epoxy group (Scheme 7.11) in the presence of a crown ethers (CE) [195–197], and in good yield. As shown in Scheme 7.11, during the ﬁrst part of the reaction an activated carbonate is formed which reacts with a second molecule of epichlorohydrin to afford the ﬁnal CC. Conversely, if the reaction were to be carried out by reacting the epihalohydrin with potassium hydrogen carbonate, a different mechanism would occur such that the 4-hydroxymethyl-1,3-dioxolan-2-one (glycerol carbonate) (Scheme 7.12) [196] would be obtained. Epichlorohydrin has been reacted with carbon dioxide (0.6 MPa) at 393 K in the presence of a zeolite-based solid catalyst TS-1, to afford cyclic carbonate in high yield [188]. The cycloaddition of CO2 to epichlorohydrin has also been performed without any solvent, in the presence of ILs as promoters. In this case, 1-alkyl-3-methyl imidazolium salts of different alkyl groups (C2, C4, C6, C8) and anions (Cl−, BF4−, Br−, PF6−) were used for the reaction, which was carried out in a batch autoclave reactor. The conversion of epichlorohydrin was seen

190

7 Synthesis of Linear and Cyclic Carbonates

O X O

18-crown-6

+ K2CO3

K+-O

O

X - +

OK O K+ O-

K+OO

X O-K+

O

O K+O-

O O

O O

X O

O

O

X

O

O

O

- +

O

OK

O

O

O Scheme 7.11 of CE.

Synthesis of cyclic carbonates from epihalohydrins and K2CO3 in the presence

O X O

+ KHCO3

18-crown-6

HO

O

X - +

OK

HO

O O

HO

X O -K +

O O O

Scheme 7.12 Synthesis of cyclic carbonates from epihalohydrins and KHCO3 in the presence of a crown ether.

to increase with the temperature, from 233 to 313 K, and with increasing CO2 pressure [198]. 7.3.3.2 From Halogenated Carbonates Cyclic carbonates have been obtained in 82–90% yield by heating monohalogenated linear organic carbonates at 353–373 K for 1–4 h (Equation 7.19) [199]. This process involved an internal nucleophilic attack of the carbonate onto the alicyclic

7.3 Synthesis of Organic Cyclic Carbonates

halide group so as to displace the chloride; the Cl− anion then reacted with the cyclic cationic intermediate to produce an alkyl chloride as the byproduct. CH3 Cl CH2 CH O C OC2H5

Hg(OAc)2 O

353 K

O

O + C2H5Cl

(7.19)

O

Similar results have been obtained starting from di-halogenated carbonates [200]. 7.3.3.3 Reaction of Cyclic Ketals with Carbon Dioxide The reaction of cyclic ketals with CO2 (Equation 7.20) under supercritical conditions in organic solvents led to the production of a CC, under relatively mild conditions (10 MPa, 370 K) and using a suitable catalyst [201].

O cat

O + SC-CO 2

O

C

O

O +

(7.20)

O The coproduct cyclohexanone may react with 1,2-ethane-diol in the presence of FeCl3 to produce, in almost quantitative yield, the cyclic ketal (Equation 7.21), which can be reused. O O

HO +

OH

Fe(III)

O

(7.21) + H2O

As a consequence, several metal systems were tested, either oxides [ZnO, Nb2O5, ZrO2, TiO2] or metal halides [ZnCl2, FeCl2], or else metal complexes [FeCl2 1.5 THF], CuL2, and FeClL. The most active catalysts were found to be CuL2 and FeClL (L = C11H7F4O2) – that is, those bearing perﬂuoroalkyl groups which were soluble in scCO2 under the reaction conditions. 7.3.4 Synthesis of Cyclic Carbonates from Propargylic Alcohols

Cyclic carbonates have also been synthesized from propargylic alcohol derivatives and CO2 as the starting materials. This synthetic approach (Equation 7.22) is based on cyclization of the propargylic carbonate moiety (HC≡CCH2OCO2–) into the corresponding α-alkylidene CC, in the presence of a suitable catalyst such as ruthenium [202], cobalt [203], palladium [204, 205], copper [206–211], or phosphine [212–214].

191

192

7 Synthesis of Linear and Cyclic Carbonates

R2 R

R2 R

R1 2

R1

+ CO2

cat

O

2

(7.22)

O

OH O The reaction usually proceeds in volatile organic solvents, such as DMF or THF, and requires large amounts of organic solvents, and a high CO2 pressure (ca. 5.0 MPa). Recently, Ikarya has reported the use of imidazolin-2-ylidenes with N-alkyl and N-aryl substituents and their CO2 adducts as catalyst of the carboxylative cyclization of internal and terminal propargylic alcohols [215]. The reaction of internal propargyl alcohols with CO2 has been carried out also under supercritical conditions. Ikariya et al. have developed a synthetic process to afford Z-alkylidene cyclic carbonates promoted by P(n-C4H9)3 with high efﬁciency [216]. An IL (1-butyl-3-methylimidazolium benzene sulfonate; [BMIm][PhSO3]) has also been used as a reaction medium for the synthesis of α-methylene CCs from CO2 and propargyl alcohols, using transition metal salts as the catalyst (Equation 7.23) [217].

OH

+ CO2

[BMIm][PhSO3] cat, 293 K

O

O

(7.23)

O The conversion rates, and the selectivities of the various catalysts used, are listed in Table 7.2. Among these, CuCl was seen to be the most efﬁcient, but no carbonate was produced; however, the substrate was very effectively converted when noble metal salts such as Pd(II), Rh(III), Ru(III), and Au(III) were used as catalysts. This behavior may be explained only if a polymerization reaction were to occur when the noble metal salts/[BMIm][PhSO3] systems had been used, as conﬁrmed by the presence of some black tar on the inner wall of the reactor when the reaction had been completed. In the absence of a metal salt as catalyst, the reaction did not yield any product, even after a long reaction time. 1,3-Dimethylimidazolium-2-carboxylate and 1-butyl-3-methylimidazolium-2carboxylate [218] have also been used as catalysts of the carboxylative cyclization of terminal propargyl alcohols, with yields ranging from 55% to 77% under mild conditions (6 MPa CO2, 273 K, 15 h). 7.3.5 Reaction Between Carbon Dioxide and Diols

Cyclic carbonates can be produced from diols and CO2 in the presence of suitable catalysts (Equation 7.24).

7.3 Synthesis of Organic Cyclic Carbonates Table 7.2

Catalysts used, and their activities.

Catalyst

Conversion of alcohol (%)

Yield (%) of carbonate

Selectivity of carbonate (%)

CuCl CuBr CuI CuCl2 FeCl3 FeCl2 CoCl2 Co(OAc)2 ZnCl2 Ni(OAc)2 PdCl2 Pd(OAc)2 RuCl3 RhCl3 No catalyst

99 99 98 84 12 10 5 3 0 2 >99 >99 >99 >99 0

97 96 97 82 – – – – – – – – – – –

>99 >99 >99 >99 99 >99 >99 0 0 99 0 0 0 0 0

R

OH R

OH

+ CO2

cat O

O

+ H 2O

(7.24)

O The thermodynamics of this reaction were not very favorable, and the major drawback related to the coproduction of water that may have involved the modiﬁcation or deactivation of the catalyst, with negative effects on the conversion rate. Ceria-based catalysts [219] and CeO2 ZrO2 solid solution catalysts [220] have each been reported to be very efﬁcient catalysts for the synthesis of EC and PC by reaction of CO2 with ethene glycol and propene glycol, respectively. The catalytic activity has been shown to depend heavily on the composition and calcination temperature of the catalysts. Different metallic acetates [221] have also been used in acetonitrile, which acts not only as a solvent but also as a dehydrating agent to eliminate the effect of any water produced during the reaction. In this way, the thermodynamic equilibrium could be shifted and the yield of CCs improved. By using 1,2-propene glycol as the reactant (100 mmol) and anhydrous zinc acetate (2.5 mmol) as catalyst in acetonitrile (10 ml) with a CO2 reaction pressure of 10 MPa, at a reaction temperature of 343 K and a reaction time of 12 h, the yield of 1,2-propene carbonate was shown to be 24.2% and the conversion of 1,2-propene glycol 38.9%. Organic “super bases,” such as DBU, or 1,5-diazabicyclo[4.3.0]non-5-ene (DBN) have also been used as effective promoters in the synthesis of PC from propene

193

194

7 Synthesis of Linear and Cyclic Carbonates

glycol and CO2 in the presence of acetonitrile. When using 1,5,7-triazabicyclo[4.4.0] dec-5-ene (TBD) under optimal conditions, the yield of PC was reported as 15.3%, with a selectivity of 100% [222]. Recently, it has been reported that magnesium and its oxide have been used as catalysts for the highly selective synthesis of carbonate through the carbonylation of a variety of 1,2-diols such as glycol, phenyl glycol with CO2, without any organic solvents or additives [223]. In the presence of 0.05 mmol% of Mg or MgO, a TON > 20 was achieved, with 100% selectivity. 7.3.6 Reaction of Urea and Diols

The reaction of urea with alkylene glycol offers not only a simple and sustainable route to the synthesis of CCs, but also a positive economic impact on the production of DMC from EC or PC, while producing large amounts of glycol as a byproduct (Scheme 7.13). The byproducts ethene glycol (EG) or propene glycol (PG) may be reacted with urea to produce again EC or PC, which is used as a starting material for the synthesis of DMC. The released ammonia can be recycled to produce urea by reaction with CO2 (Scheme 7.13). It was ﬁrst reported by Su and Speranza [224] that a tin-based catalyst would be active in such a reaction. In fact, the conversion of PG was reported as 43%, and the yield and selectivity of PC as 36% and 84%, respectively, although a severe loss of ammonia that could not be recycled was also observed. Later, Yutaka et al. [225] and Doya et al. [226] patented a new process for the production of PC from PG and urea under reduced pressure, using a catalyst that

CO2 O H2N

2NH3 NH2 cat. OH

H3C

O OH

H3CO

O

CH3OH

O

Scheme 7.13

O

H 3C

OCH3

Reaction of urea with alkylene glycol relevant to the synthesis of DMC.

7.3 Synthesis of Organic Cyclic Carbonates

contained at least one metal selected from among zinc, magnesium, lead, and calcium; in this case the yield of PC was up to 97.2%. However, when using this method the extra consumption of energy was required for assembling the vacuum equipment. Other catalytic systems to be used, such as zinc acetate and supported zinc acetate [227], reached PC yields of 94% and 78%, respectively. Notably, a serious loss of zinc acetate was observed with the supported catalyst. Several metal oxides (either acidic or alkaline) have also been investigated for urea alcoholysis [228, 229], with PG ﬁnding PC product yields in excess of 90% for ZnO, PbO, and MgO. In such studies, the results obtained coupled with the results of thermal programmed desorption (TPD) and Fourier transform infrared (FTIR) analyses, indicated that catalysts with appropriate acid and base properties were required for the synthesis of CCs. These results conﬁrmed the reports of Aresta et al. [94] and Ball et al. [39], who previously had investigated the reaction of primary and secondary alcohols with urea to form carbonate. These authors found the reaction to proceed in two steps, with a combination of a weak Lewis acid and a Lewis base improving the carbonate formation. The production of PC from urea and 1,2-propanediol has also been performed, in a batch process, using zinc chloride and magnesium chloride [230]. Under optimal reaction conditions (ethanol : urea molar ratio 4, catalyst concentration 1.5%, reaction temperature 333 K, reaction time 3 h), both MgCl2 and ZnCl2 showed excellent catalytic activity towards PC synthesis, with the yields reaching 96.5% and 92.4%, respectively. 7.3.7 Reaction of Carbon Dioxide or Urea with Glycerol

Glycerol carbonate can be produced by reacting glycerol with phosgene, or with: (i) a dialkyl carbonate [231, 232]; (ii) an alkylene carbonate [233–235]; or else by the reaction of glycerol with urea [236–238], or by reaction with carbon monoxide and oxygen in the presence of Cu(I) catalysts [239]. Glycerol does not react with CO2 (Equation 7.25) in the presence of zeolites as catalyst [240], whereas Sn-catalysts [n-Bu2Sn(OMe)2, n-Bu2SnO, Sn(OMe)2] promote the carboxylation of glycerol [241] at 5 MPa and 450 K.

O

H H

OH

H

OH

H

OH H

+

CO2

cat.

O O

+

H2O

CH2OH (7.25)

195

196

7 Synthesis of Linear and Cyclic Carbonates Table 7.3 Carboxylation of glycerol using di (n-butyl)tindimethoxide as catalyst.a

% catalyst

Time (h)

P CO2

Temperature

Solvent

Molecular sieves

Glycerol conversion % (isolated)b

2 2 2 2 2 6 6 6

15 15 15 15 15 6 10 15

5 MPa 5 MPa 2.5 MPa 5 MPa 5 MPa 5 MPa 5 MPa 5 MPa

453 K 453 K 453 K 373 K 453 K 453 K 453 K 453 K

None None None None Tedmgc None None None

No Yes Yes Yes Yes Yes Yes Yes

0.42 (0.35) 2.29 (1.94) 1.74 (1.49) 0.49 (0.42) 2.22 (1.88) 3.30 (2.80) 5.87 (4.93) 6.86 (5.72)

a) In all experiments 4 g of glycerol (43.5 mmol) and 0.78 g of catalyst (2.61 mmol) were used. b) The reaction yield was in all cases 15–20% higher than the isolated yield. c) tedmg = tetraethylene glycol dimethyl ether.

The most active among the catalysts tested was n-Bu2Sn(OMe)2; the results obtained under different reaction conditions are listed in Table 7.3. The reaction was carried out at 450 K in glycerol or in tedmg as solvent under 5 MPa of CO2. Glycerol carbonate was formed with an appreciable rate until a 1.14 : 1 molar ratio of carbonate with the catalyst was reached. The monomeric species formed from the reaction of n-Bu2Sn(OMe)2 with glycerol was able to react with CO2 until it was a monomer. During the reaction, the original catalyst was converted into an oligomer which showed a moderate catalytic activity, thus explaining the low TOF that was encountered. When the mechanism of the reaction was elucidated, n-Bu2Sn(OCH3)2 was shown to react with glycerol at room temperature in toluene with the elimination of methanol to afford a n-Bu2Sn(glycerol). This then reacted with CO2 only at high temperature under 5 MPa of CO2, to form the carboxylated complex, as demonstrated by an FTIR study under pressure (νCO at 1681 cm−1). The experimental data obtained suggest that once the monomeric species n-Bu2Sn(glycerol) was formed, it could either incorporate CO2 or oligomerize with deactivation. Once formed, the carboxylated species was able to eliminate glycerol carbonate, affording a Sn-compound characterized by Sn–O–Sn bonds similar to that reported by Ballivet-Tkatchenko [80, 88], when the same catalyst was used in the carboxylation of methanol. These results demonstrate that, when the oligomer is formed, both the reactivity of the complex towards CO2 and the activity of the catalyst are slowed down. Subsequently, the mechanism shown in Scheme 7.14 was demonstrated experimentally [241]. The glycerolysis of urea represents an alternative synthetic approach for the synthesis of glycerol carbonate (Equation 7.26):

7.3 Synthesis of Organic Cyclic Carbonates

197

OH

O

OH

R

R

OMe OH

OMe

R

CO2

Sn

Sn R

O

R

- 2 MeOH

O

O O

O

Sn

O

O

R

OH

HO

O

+ [R2SnO]n

HO

Scheme 7.14 Reaction mechanism of formation of glycerol carbonate from glycerol and CO2 under n-Bu2Sn(OMe)2 catalysis.

O

H H

OH

H

OH

H

OH

O

cat.

O

+

O H2N

+

2 NH3

NH2

H

CH2OH

(7.26) Here, the reaction proceeds quickly enough simply by heating the mixture of urea and glycerol, although the addition of a catalytic system will improve the conversion yield of glycerol. The role of the metal center is to facilitate the interaction between the urea carbonyl and the glycerol alcoholic group, with the consequent release of ammonia. The formation of glycerol carbonate takes place in two consecutive steps, the ﬁrst of which is formation of the carbamate species (Equation 7.27), with removal of the ﬁrst molecule of ammonia followed by the formation of carbonate in the second step and contemporary elimination of the second molecule of ammonia (Equation 7.28).

H H

OH

H

OH

H

OH

O

O +

H2N

NH2

H2N

OCH2CH(OH)CH2OH + NH3

H (7.27)

CH2OH

O H2N

OCH2CH(OH)CH2OH

O

O

+ NH3

O (7.28)

198

7 Synthesis of Linear and Cyclic Carbonates Table 7.4 Glycerolysis of urea using several different catalytic systems.a

Catalyst

n glycerol/n urea

w catalyst/w urea (%)

Recoverability

Conversion (%)

None Alloy Al–Ce–Ga TiO2 CeO2 Rh(diphos)BPh4 Bu2SnO Titanosilicalite Titanosilicalite Bi2O3 ZnO γ-ZrP Zn γ-ZrP not calcined γ-ZrP not calcined γ-ZrP calcined

2 2 2 2 2 2 1 2 2 2 2 1 2 1

– 5 5 5 3 5 3 3 5 3 3 1 1 1

– Yes No No Yes No Partial Partial No No Yes Yes Yes Yes

28 30 32 32 35 36 36 58 42 48 62 60 68 76

a) The conversion is referred to urea, as glycerol was used as solvent. Conditions: 413 K, 10−3 bar, 3 h.

The role of the catalyst is very important in the latter step, which requires hard reaction conditions (M. Aresta et al. unpublished results). In particular, by using metal oxides characterized by a speciﬁc ratio between the acid and basic sites [242] as catalyst, conversion of the carbamate into the carbonate is very much improved, together with selectivity (the conversion into carbonate increases as the acid/basic site ratio decreases). Aresta et al. have investigated several catalytic systems, under various reaction conditions, and have patented a method [235] of recovering the catalyst and isolating and purifying the product. For this, different metal systems were used (Table 7.4) in an attempt to identify the most efﬁcient system, by considering not only the metal’s activity but also its recoverability and reuse. Based on the data collected, the most active catalyst was found to be γ-ZrP. Although the conversion yield of glycerol into glycerol carbonate is comparable with that obtained in other previously patented processes, where anhydrous MgSO4 [243, 244] or calcined ZnSO4 [245] were used as catalyst, the use of γ-ZrP seemed more effective as the catalyst could be very easily recovered from the reaction solution and reused for several cycles. Moreover, the catalyst maintained the same activity and selectivity for several cycles if it was recovered and calcined after two to three cycles. 7.3.8 Reactivity of Cyclic Alkylene Carbonates

Previously, CCs have found a variety of applications in organic syntheses; some of those related to EC are reported in Scheme 7.15.

7.3 Synthesis of Organic Cyclic Carbonates

199

R3N-EC R2NHC(O)OCH2CH2OH

R3N

rN A

ArN(CH2CH2OH)

NH

H2

R2

(-OC-C6H4-C O(O)CH2CH2O-)n

R NH

PhAnhy

O

2

RNHC(O)OCH2CH2OH

O RSH

O

RC

Diols H(OROCO)xOROH

ROH

RSHCH2CH2OH OO H

RC(O)OCH2CH2OH

1. RO-C(O)OCH2CH2OH (transest. cat) 2. RO-C(O)OR + HOCH2CH2OH (transest. cat with excess ROH) 3. ROCH2CH2OH (basic cat.) Scheme 7.15 Ethene carbonate applications.

R R1

N

+

R

O

H O

O

R2 O

R1

N R2

OH

O R

Δ -H2O

N

O

R1

R

Δ R1NH2 -H2O

O

Scheme 7.16 Synthesis of oxazolidinones and imidazolidinones.

For example, EC yields polymers containing blocks of poly-ethers and polycarbonates. due to the partial elimination of CO2 [246], while its hydrolysis produces high-purity 1,2-diols. Furthermore, CCs react readily with carboxylic acids to form 2-hydroxyethyl esters [177, 247, 248]. Alkylene carbonates can be used as alkylating agents (in the presence of suitable catalysts) of active-hydrogen-containing aromatics such as phenols [249, 250], thiophenols [251–253], aniline [254], and carboxylic acids [255, 256]. They may react with aliphatic amines undergoing an attack at the carbonyl carbon atom, followed by ring opening to produce urethanes [257] which, in the absence of amines and upon loss of water, generate oxazolidinones. The latter may react in the presence of an excess of amines to afford imidazolidinones [258] (Scheme 7.16). Recently, a method for the synthesis of 2-oxazolidinones and 2-imidazolidinones from ﬁve-membered CCs and β-amino alcohols or 1,2-diamines using Br−Ph3+PPEG600-P+Ph3Br− as a homogeneous recyclable catalyst has been described [259] (Equation 7.29).

N

N

R1

R1 O

200

7 Synthesis of Linear and Cyclic Carbonates

R

O O + R1

O

HO

OH

N H

N

Scheme 7.17

O

R

R1

O

OH

O

NR1

HO

OH

+ R

O

Synthesis of oxazolidinones using hydroxoalkylamines.

R

R HX

O

O

+

NH2

cat

X

HO

NH

+

R'

O

OH

R'

O

X=O, NH (7.29) Alkylene carbonate may also react with a hydroxoalkylamine to create a cyclization reaction [260] that affords oxazolidinones (Scheme 7.17). The ring-opening polymerization does not occur easily; in fact, many reports have indicated that the ring-opening process requires an anionic species, followed by propagation during which an ether linkage may be formed due to the loss of CO2 [261]. The cyclic alkylene carbonates have also been applied as the cure-accelerators of phenol–formaldehyde (PF) [262–266] and sodium silicate [267] resin systems, which are widely used in foundry sand and wood binder applications.

7.4 Transesteriﬁcation Reactions 7.4.1 Synthesis of Acyclic Carbonates

Very often, the transesteriﬁcation reaction implies the need for an alkylene or CC (e.g., EC or PC) and an alcohol in the presence of either a homogeneous or heterogeneous acidic or basic catalyst [268], to co-produce dialkyl carbonate and the alkane diol or glycol (Equation 7.30).

R

O

O

O O

cat + 2 ROH

R

O

O

R

HO

OH

+ R

R = H, CH3 (7.30)

7.4 Transesteriﬁcation Reactions

O

+ CO2

cat

O

O

+ 2CH 3OH

cat

HO

OH +

201

O CH3O C OCH3

O Scheme 7.18 Synthesis of DMC starting from epoxide.

Subsequently, a number of homogeneous catalysts have been patented and described as capable of promoting such reactions; these include tertiary amines [269–272], and zirconium, titanium and tin complexes [273]. Recently, Feng [274] reported a continuous process for the transesteriﬁcation of EC with methanol in a ﬂow reactor over a dibutyl amine catalyst immobilized on a MCM-41 molecular sieve (n-Bu2N-MCM-41). The catalyst performed well and afforded 25.5% and 41.7% EC conversions at 283 and 323 K, respectively, and also exhibited a good stability. Arai and coworkers [275, 276] subsequently performed the preparation of DMC in two steps (Scheme 7.18). Initially, EC was prepared by reacting the relevant epoxide with CO2, while the latter was reacted with methanol using MgO, CaO, ZnO, ZrO2, La2O3, CeO2, and Al2O3 as catalysts. MgO was reported to be the best catalyst for both reactions, as it gave the highest conversion (35%) and selectivity (92%). Although the one-pot synthesis of DMC is possible, the selectivity is not good due to alcoholysis of the epoxide [277]. In addition, CaO showed a unique catalytic activity for the transesteriﬁcation of PC with methanol, with a high PC conversion and DMC selectivity [278]. However, when CaO was used in a ﬂow reactor to produce DMC, a reduction in its activity was noted due to the leaching of calcium [279]. With regards to supported catalysts, much effort has been expended in attempts to prepare stable anchored active species on carriers, the aim being to overcome the problem of metal leaching. For example, active species have been grafted onto mesoporous materials, encapsulated using “ship-in-a-bottle” techniques, and ion-exchanged in layered structures [280]. CaO–ZrO2 solid solutions have been used to prevent the leaching of calcium via the strong interaction between CaO and ZrO2 in solid solution [281]. Under optimal reaction conditions, CaO–ZrO2 showed a high stability towards the transesteriﬁcation of PC with methanol (see Figure 7.2). Notably, the conversion of PC was maintained at about 95% for up to 250 h. Sun and colleagues [282] have also investigated the effect of the base strength of MgO and CaO, with both materials showing good catalytic activity in the synthesis of DMC via the transesteriﬁcation of PC with methanol. Here, both the base strength and basicity were found to have signiﬁcant inﬂuences on the catalytic activity – that is, the higher the base strength of a catalyst, the lower was the temperature needed. In contrast, the increase in basicity resulted in a rise in reaction rate, with a slight decrease in selectivity. The use of catalysts such as KOH, NaOH, K2CO3 and KNO3, supported on molecular sieves 4 Å, was ﬁrst studied by Li [283] when investigating the synthesis of DMC from methanol and PC synthesized from PO and CO2. Based on the

7 Synthesis of Linear and Cyclic Carbonates 105 100 95 PC Coversion/%

202

90 85 80 75 70 65 60 55 50 0

Figure 7.2

50

100 150 Reaction Time/h

200

250

Stability of the CaO–ZrO2 catalyst.

results of these studies, it was concluded that the KOH/4 Å molecular sieve was the best catalyst, and could be efﬁciently recycled. Huang [284] has reported details of the synthesis of DMC using KI, K2CO3, KOH, and NaOH supported on MgO or ZnO as catalysts in scCO2. Both the KI/ZnO and K2CO3–KI/ZnO combinations were shown to produce good catalysts, with complete conversion of epoxides being achieved in 4 h, with a high selectivity. The same reaction was also investigated by Cui and coworkers [285], who showed that K2CO3, when used as catalyst, could provide 51.7% DMC selectivity and 36.3% EC selectivity when the reaction was carried out at 293 K, with 15.0 MPa of CO2 pressure. However, the K2CO3–KI/ZnO combination proved to be the best catalyst system. The effect of scCO2 on the selectivity and conversion of the transesteriﬁcation of EC with methanol, using K2CO3 as catalyst, has been investigated at pressures of up to 30 MPa and temperatures ranging from 273 to 313 K [286]. Subsequently, a high CO2 pressure was found to suppress not only the transesteriﬁcation of EC with methanol catalyzed by K2CO3, but also – and more signiﬁcantly so – the undesired side reactions. In fact, at a high CO2 pressure the EC conversion and DMC yield reached 47.9% and 47.0%, respectively, with a DMC selectivity of 98.1% at 313 K and 15.0 MPa with a MeOH : EC molar ratio of 4 and 2.5 wt% K2CO3 based on EC–MeOH solution for a 1.5 h reaction time. With scCO2 as the medium, the transesteriﬁcation reaction could be carried out at higher temperatures so as to obtain both a high EC conversion and a high selectivity. When Mg–Al–CO3 hydrotalcite was tested for the synthesis of DMC via transesteriﬁcation, this heterogeneous basic catalyst showed good activity for the process [287]. Cu–KF/MgSiO was also reported to be capable of catalyzing such a process [84]. Metal alkoxides such as V-, Nb- and Ta-(V) alkoxides, as well as oxides in various oxidation states – in particular Nb(II, III, IV, and V), V(III, IV, and V) and Ti (IV)oxides – have been studied in transesteriﬁcation reactions involving EC and

7.4 Transesteriﬁcation Reactions

DMC formation / %

80 60 40 20 0 0

Figure 7.3

30 10 Time / h 20 Ta(OMe)5 Nb(OMe)5 VO(OiPr)3 NbO(OEt)3 Proﬁle of the transesteriﬁcation reaction of ethene carbonate with methanol.

Conversion of Carbonate / %

100 80 60 40 20 0 0

10

V2O5 (DEC) V2O3 (DMC)

Time / h

20

V2O5 (DMC) V2O4 (DMC)

30

TiO2 (DMC) NbO2 (DMC)

Figure 7.4 Catalytic performance of metal oxides used in the transesteriﬁcation of ethene carbonate with methanol.

aliphatic C1–C3 alcohols [288]. In the case of the metal alkoxides, the VO(OiPr)3 and Nb-alkoxides were shown to be more active than both the [NbO(OEt)3]2 and Ta-alkoxides (Figure 7.3). Notably, the TOF during the ﬁrst 4 h was respectively 1 mol of DMC per mol catalyst per h for the V complex, 0.9 for [Nb(OMe)5]2, and only 0.1 for the [NbO(OMe)3]2 and Ta-system. The data in Figure 7.3 show that an equilibrium is reached which corresponds to a conversion of 80% of the initial carbonate in the case of the V- and Nb-catalysts. Moreover, the reaction can be continued for many hours, without any signiﬁcant change in the conversion, although with Ta the reaction is much slower. The best performance observed was with VO(OiPr)3, which demonstrated a TON of 5.22. The catalysts each maintained their catalytic activity for several cycles, and could be isolated and reloaded without losing their activity. When the tested oxides were tested, they showed a catalytic activity that, on occasion, was comparable with that of the alkoxides. The reaction proﬁle using the various oxides as catalysts is shown in Figure 7.4. The best performance was

203

204

7 Synthesis of Linear and Cyclic Carbonates

Figure 7.5 Numbers of patents (black area) and publications (white area) in the years 1972–2008. (a) Processes using DPC; (b) DPC synthesis from DMC. CAS database.

provided by V2O5, which was able to convert EC at a level of 60% after 3 h, with a TOF equal to 0.95 h−1. Although V2O5 was shown to be as effective as TiO2, unfortunately both oxides were able to react with the alcohol, affording an oxidation of alcohol to aldehyde with catalyst reduction and deactivation. In fact, as shown in Figure 7.4, V2O3 and V2O4 were worse catalysts than V2O5. Among the Nb-oxides, NbO2 was the most active. The number of patents detailing processes that involve DPC as an intermediate to polycarbonates, polyurethanes and organics bears witness to the importance of this molecule in the production of industrial polymers and molecules (Figure 7.5a). The data in this ﬁgure also underline the recent surge in open-literature reports, a trend which was also apparent for the synthesis of DPC from DMC (Figure 7.5b). 7.4.2 Synthesis of Cyclic Carbonates

Alkylene CCs have been prepared through the transesteriﬁcation of appropriate glycols with dialkyl carbonates (usually diethyl or dimethyl carbonate) in the presence of a suitable catalyst. One of the ﬁrst such examples was the synthesis of six-membered CCs by the transesteriﬁcation of propane-1,3-diols with DEC catalyzed by sodium ethanolate (Equation 7.31) [289]. The reaction was carried out at temperatures between 293 and 333 K, and a conversion yield of 40% was obtained.

HO

OH

+

Et

O

O O

Et

cat O

O O

(7.31)

References

A similar method has been reported by Albertsson et al. [290], in which equimolar amounts of propane-1,3-diol and DEC were used, with stannous 2-ethylhexanoate as the transesteriﬁcation catalyst, affording a yield of 53%. Other examples [291, 292] have included the use of propane-1,3-diols which had been differently substituted and treated with DEC in the presence of catalytic amounts of sodium methoxide. Then, depending on the reaction conditions, either polycarbonates or CCs were produced in high yield. Six-membered alkylene carbonates have also been synthesized by reacting several 1,3-diols with a 15% excess of EC in the presence of titanium(IV) isopropoxide at 293–323 K and 15–30 mmHg [293]. In this case, the high-purity (99%) alkylene carbonate was obtained via a short-path distillation at 323–373 K at reduced pressure. Alkylene carbonates have also been obtained by the disproportionation of 1,3-bis(alkoxycarbonyloxy)propanes, using colloidal silica or Sn(II) stearate [294]. The transesteriﬁcation of 1,2-diols by reaction with carbonates, both cyclic and linear, produces ﬁve-membered alkylene carbonates almost exclusively. A wellknown example of this is the reaction of DMC with propene glycol to yield PC [295].

References 1 Aresta, M. and Galatola, M. (1999) J. Cleaner Prod., 7, 187–193. 2 Shaikh, A.-A.G. and Sivaram, S. (1996) Chem. Rev., 96, 951. 3 Parrish, J.P., Salvatore, R.N., and Jung, K.W. (2000) Tetrahedron, 56, 8207. 4 Pensado, A.S., Padua, A.A.H., Comunas, M.J.P., and Fernandez, J. (2008) J. Phys. Chem. B, 112, 5563. 5 Wang, Y., Nakamura, S., Tasaki, K., and Balbuena, P.B. (2002) J. Am. Chem. Soc., 124, 4408. 6 Li, Z., Shan, F., Wei, J., Yang, J., Li, X., and Wang, X. (2008) J. Solid State Electrochem., 12, 1629. 7 Machac, J.R., Jr, Marquis, E.T., and Woodrum, S.A. (2000) US Patent 654438. 8 Deswartvaegher, A., Forestier, B., Miard, S., Senet, J.P., Thiebaud, R.S., Cristea, D., De Caro, P., and Giacinti, G. (2006) WO 056558. 9 Behr, A., Bahke, P., Klinger, B., and Becker, M. (2007) J. Mol. Catal. A: Chemical, 267, 149. 10 Bernini, R., Mincione, E., Barontini, M., Crisante, F., Fabrizi, G., and Gambacorta, A. (2007) Tetrahedron, 63, 6895.

11 Schäffner, B., Holz, J., Verevkin, S.P., and Börner, A. (2008) ChemSusChem, 1, 249. 12 Pacheco, M.A. and Marshall, C.L. (1997) Energy Fuels. 11, 2. 13 Li, X., Chen, H., Zhu, Z., and Zhen, H. (2006) Energy Conv. Mgmt, 47, 1438. 14 Li, D., Fang, W., Xing, Y., Guo, Y., and Lin, R. (2009) J. Hazard. Mater., 161, 1193. 15 Cicha, W.V. and Manzer, L.E. (1997) US Patent 22903. 16 Xiangyang, Y. (2002) China is making a vigorous development of DMC. China Chemical Reporter, October 16, p. 16. 17 Rivetti, F. (2000) C. R. Acad. Sci., Ser. IIc: Chim., 3, 497. 18 Katrib, Y., Deiber, G., Mirabel, P., Le Calvé, S., George, C., Mellouki, A., and Le Bras, G. (2002) J. Atmos. Chem., 43, 151. 19 Anastas, P.T. and Kirchhoff, M.M. (2002) Acc. Chem. Res., 35, 686. 20 Anastas, P.T. and Zimmerman, J.B. (2003) Environ. Sci. Technol., 37, 95A. 21 Ono, Y. (1997) Appl. Catal. A, 155, 133. 22 Tundo, P. and Selva, M. (2002) Acc. Chem. Res., 35, 706.

205

206

7 Synthesis of Linear and Cyclic Carbonates 23 Selva, M. and Perosa, A. (2008) Green Chem., 10, 457. 24 Romano, U., Tesei, R., Mauri, M.M., and Rebora, P. (1980) Ind. Eng. Chem. Prod. Res. Dev., 19, 396. 25 Delledonne, D., Rivetti, F., and Romano, U. (2001) Appl. Catal. A, 221, 241. 26 Uchiumi, S.-I., Ataka, K., and Matsuzaki, T. (1999) J. Organomet. Chem., 576, 279. 27 Fukuoka, S., Kawamura, M., Komiya, K., Tojo, M., Hachiya, H., Hasegawa, K., Aminaka, M., Okamoto, H., Fukawa, I., and Konno, S. (2003) Green Chem., 5, 497. 28 Fukuoka, S., Tojo, M., Hachiya, H., Aminaka, M., and Hasegawa, K. (2007) Polym. J., 39, 91. 29 Trost, B.M. (2002) Acc. Chem. Res., 35, 695. 30 Malone, M.F., Huss, R.S., and Doherty, M.F. (2003) Environ. Sci. Technol., 37, 5325. 31 Thotla, S. and Mahajani, S.M. (2009) Chem. Eng. Res. Des., 87, 61. 32 Wang, F., Zhao, N., Li, J., Zhao, W., Xiao, F., Wei, W., and Sun, Y. (2007) Ind. Eng. Chem. Res., 46, 8972. 33 Saleh, R.Y., Michaelson, R.C., Suciu, E.N., and Kuhlmann, B. (1994) US Patent 14540. 34 Ryu, J.Y. (1998) US Patent 140435. 35 Lin, H., Yang, B., Sun, J., Wang, X., and Wang, D. (2004) Chem. Eng. J., 103, 21. 36 Wang, M., Zhao, N., Wei, W., and Sun, Y. (2005) Ind. Eng. Chem. Res., 44, 7596. 37 Wang, M., Wang, H., Zhao, N., Wei, W., and Sun, Y. (2006) Catal. Commun., 7, 6. 38 Ryu, J.Y. and Gelbein, A.P. (2001) US Patent 866340. 39 Ball, P., Fullmann, H., and Heitz, W. (1980) Angew. Chem. Int. Ed. Engl., 19, 718. 40 Heitz, W., Fuellmann, H., and Ball, P. (1981) German Patent 3112794. 41 Suciu, E.N., Kuhlmann, B., Knudsen, G.A., and Michaelson, R.C. (1998) J. Organomet. Chem., 556, 41. 42 Ryu, J.Y. and Gelbein, A.P. (2004) US Patent 821225.

43 Wang, D., Yang, B., Zhai, X., and Zhou, L. (2007) Fuel Process. Technol., 88, 807. 44 Dubois, J.-L. (2008) WO 139080. 45 Cella, J.A. and Bacon, S.W. (1984) J. Org. Chem., 49, 1122. 46 Jorapur, Y.R. and Chi, D.Y. (2005) J. Org. Chem., 70, 10774. 47 McGhee, W. and Riley, D. (1995) J. Org. Chem., 60, 6205. 48 Fang, S. and Fujimoto, K. (1996) Appl. Catal. A: General, 142, L1. 49 Isaacs, N.S., O’Sullivan, B., and Verhaelen, C. (1999) Tetrahedron, 55, 11949. 50 Fujita, S.-I., Bhanage, B.M., Arai, M., and Ikushima, Y. (2001) Green Chem., 3, 87. 51 Hoffman, W.A., III (1982) J. Org. Chem., 47, 5209. 52 Aresta, M., Dibenedetto, A., Fracchiolla, E., Giannoccaro, P., Pastore, C., Papai, I., and Schubert, G. (2005) J. Org. Chem., 70, 6177. 53 Yamazaki, N., Nakahama, S., and Higashi, F. (1979) Ind. Eng. Chem. Prod. Res. Dev., 18, 249. 54 Kizlink, J. and Pastucha, I. (1994) Collect. Czech. Chem. Commun., 59, 2116. 55 Wagner, A. and Haas, W. (1994) WO 022805. 56 Cheong, M., Kim, S.-C., and Park, J.B. (1997) New J. Chem., 21, 1143. 57 Sakakura, T., Choi, J.-C., Saito, Y., Masuda, T., Sako, T., and Oriyama, T. (1999) J. Org. Chem., 64, 4506. 58 Choi, J.-C., Kohno, K., Ohshima, Y., Yasuda, H., and Sakakura, T. (2008) Catal. Commun., 9, 1630. 59 Genz, J. and Heitz, W. (1983) European Patent 100384. 60 Choi, J.-C., He, L.-N., Yasuda, H., and Sakakura, T. (2002) Green Chem., 4, 230. 61 Hou, Z., Han, B., Liu, Z., Jiang, T., and Yang, G. (2002) Green Chem., 4, 467. 62 Aresta, M., Dibenedetto, A., Nocito, F., and Pastore, C. (2008) Inorg. Chim. Acta, 361, 3215. 63 Fan, B., Zhang, J., Li, R., and Fan, W. (2008) Catal. Lett., 121, 297. 64 Kizlink, J. and Pastucha, I. (1995) Collect. Czech. Chem. Commun., 60, 687.

References 65 Kohno, K., Choi, J.-C., Ohshima, Y., Yasuda, H., and Sakakura, T. (2008) ChemSusChem, 1, 186. 66 Aresta, M., Dibenedetto, A., and Pastore, C. (2003) Inorg. Chem., 42, 3256. 67 Dibenedetto, A., Aresta, M., Giannoccaro, P., Pastore, C., Papai, I., and Schubert, G. (2006) Eur. J. Inorg. Chem., (5), 908. 68 Tomishige, K., Sakaihori, T., Ikeda, Y., and Fujimoto, K. (1999) Catal. Lett., 58, 225. 69 Ikeda, Y., Asadullah, M., Fujimoto, K., and Tomishige, K. (2001) J. Phys. Chem. B, 105, 10653. 70 Tomishige, K., Furusawa, Y., Ikeda, Y., Asadullah, M., and Fujimoto, K. (2001) Catal. Lett., 76, 71. 71 Yoshida, Y., Arai, Y., Kado, S., Kunimori, K., and Tomishige, K. (2006) Catal. Today, 115, 95. 72 Aresta, M., Dibenedetto, A., Pastore, C., Cuocci, C., Aresta, B., Cometa, S., and De Giglio, E. (2008) Catal. Today, 137, 125. 73 Tomishige, K. and Kunimori, K. (2002) Appl. Catal. A: General, 237, 103. 74 Osora, H., Kobayashi, K., Seiki, Y., Yasutake, T., Iijima, M., and Oguchi, A. (2004) WO 091778. 75 Jiang, C., Guo, Y., Wang, C., Hu, C., Wu, Y., and Wang, E. (2003) Appl. Catal. A: General, 256, 203. 76 Allaoui, L.A. and Acuissi, A. (2006) J. Mol. Catal. A: Chemical, 259, 281. 77 Wang, X.J., Xiao, M., Wang, S.J., Lu, Y.X., and Meng, Y.Z. (2007) J. Mol. Catal. A: Chemical, 278, 92. 78 Ballivet-Tkatchenko, D., Douteau, O., and Stutzmann, S. (2000) Organometallics, 19, 4563. 79 Camy, S., Pic, J.S., Badens, E., and Condoret, J.S. (2003) J. Supercrit. Fluids, 25, 19. 80 Ballivet-Tkatchenko, D., Chambrey, S., Keiski, R., Ligabue, R., Plasseraud, L., Richard, P., and Turunen, H. (2006) Catal. Today, 115, 80. 81 Pinero, R., Garcia, J., Sokolova, M., and Cocero, M.J. (2007) J. Chem. Thermodyn., 39, 536. 82 Guo, X.C., Qin, Z.F., Wang, G.F., and Wang, J.G. (2008) Chin. Chem. Lett., 19, 249.

83 Wang, L., Li, J., Lin, Y., and Chen, C. (2009) Chem. Eng. J., 146, 71. 84 Li, C.-F. and Zhong, S.-H. (2003) Catal. Today, 82, 83. 85 Zhao, T., Han, Y., and Sun, Y. (2000) Fuel Process. Technol., 62, 187. 86 Cai, Q., Lu, B., Guo, L., and Shan, Y. (2009) Catal. Commun., 10, 605. 87 Hou, Z., Han, B., Liu, Z., Jiang, T., and Yang, G. (2002) Green Chem., 4, 467. 88 Ballivet-Tkatchenko, D., Burgat, R., Chambrey, S., Plasseraud, L., and Richard, P. (2006) J. Organomet. Chem., 691, 1498. 89 Miyake, N. and Nagahara, H. (2003) WO 055840. 90 Miyake, N. and Budianto, B. (2006) Japanese Patent 176411. 91 Budianto, B. and Miyake, N. (2007) Japanese Patent 269653. 92 Shinohata, M. and Miyake, N. (2007) WO 044575. 93 Jung, K.T. and Bell, A.T. (2001) J. Catal., 204, 339. 94 Aresta, M., Dibenedetto, A., Pastore, C., Papai, I., and Schubert, G. (2006) Top. Catal., 40, 71. 95 Choi, J.-C., Sakakura, T., and Sako, T. (1999) J. Am. Chem. Soc., 121, 3793. 96 Ballivet-Tkatchenko, D., Jerphagnon, T., Ligabue, R., Plasseraud, L., and Poinsot, D. (2003) Appl. Catal. A: General, 255, 93. 97 Kohno, K., Choi, J.-C., Ohshima, Y., Yili, A., Yasuda, H., and Sakakura, T. (2008) J. Organomet. Chem., 693, 1389. 98 Ballivet-Tkatchenko, D., Chermette, H., Plasseraud, L., and Walter, O. (2006) Dalton Trans., 5167. 99 Carothers, W.H. and Hill, J.W. (1933) J. Am. Chem. Soc., 55, 5043. 100 Spanagel, E.W. and Carothers, W.H. (1935) J. Am. Chem. Soc., (5), 929. 101 Lichtenwalter, M. and Cooper, J. (1956) US Patent 2773070. 102 McClellan, P. (1959) US Patent 2873282. 103 Sun, Y., Liu, J., Kimbleton, E., and Wang, J. (1999) US Patent 5993787. 104 Marquis, T., Baldwin, R., Machac, J., Darragas, K., and Woodrum, S. (2001) US Patent 6169061.

207

208

7 Synthesis of Linear and Cyclic Carbonates 105 Wakihara, M. and Yamamoto, O. (1998) Lithium Ion Batteries, Kodansha Ltd., Tokyo. 106 Aresta, M. and Quaranta, E. (1997) Chem. Tech., 27 (3), 32. 107 Cant, N.W. and Hall, W.K. (1978) J. Catal., 52, 81. 108 Clerici, M.G., Belussi, G., and Romano, U. (1991) J. Catal., 129, 159. 109 Clerici, M.G. and Ingallina, P. (1998) Catal. Today, 41, 351 (and refs therein). 110 Hayashi, T., Tanaka, K., and Haruta, M. (1998) J. Catal., 178, 566. 111 Uphade, B.S., Tsubota, S., Hayashi, T., and Haruta, M. (1998) Chem. Lett., 27, 1277. 112 Duma, V. and Honicke, D. (2000) J. Catal., 191, 93. 113 Stangland, E.E., Stavens, K.B., Andres, R.P., and Delgass, W.N. (2000) J. Catal., 191, 332. 114 Jenzer, G., Mallat, T., Maciejewski, M., Eigenmann, F., and Baiker, A. (2001) Appl. Catal. A: General, 208, 125 (and refs therein). 115 Aresta, M., Dibenedetto, A., Nocito, F., Pastore, C., Venezia, A.M., Chirykalova, E., Kononenko, V.I., Shevchenko, V.G., and Chupova, I.A. (2006) Catal. Today, 115, 117. 116 Dibenedetto, A., Aresta, M., Distaso, M., Pastore, C., Venezia, A.M., Liu, C.J., and Zhang, M. (2008) Catal. Today, 137, 44. 117 Ainsworth, S. (1992) J. Chem. Eng. News, 70 (9), 9. 118 Weissermel, K. and Arpe, H.J. (1997) Industrial Organic Chemistry, 3rd ed., VCH, Weinheim, p. 162. 119 Darensbourg, D.J., Andreatta, J.R., and Moncada, A.I. (2010) in Carbon Dioxide as Chemical Feedstock (ed. M. Aresta), ch. 8, Wiley-VCH Verlag GmbH, Weinheim. 120 Limura, N., Takagi, M., Iwane, H., and Ookago, J. (1995) Jpn. Kokai Tokkyo Koho, Japanese Patent 07267944. 121 Inoe, K. and Ookubo, H. (1995) Jpn. Kokai Tokkyo Koho, Japanese Patent 07206846. 122 Inoe, K. and Ookubo, H. (1995) Jpn. Kokai Tokkyo Koho, Japanese Patent 07206847.

123 Inoe, K. and Ookubo, H. (1995) Jpn. Kokai Tokkyo Koho, Japanese Patent 07206848. 124 Inaba, M., Hasegawa, K., and Nagaoka, H. (1997) Jpn. Kokai Tokkyo Koho, Japanese Patent 09067365. 125 Ichikawa, S. and Iwane, H. (1997) Jpn. Kokai Tokkyo Koho, Japanese Patent 09235252. 126 Tojo, M. and Fukuoka, S. (1991) Asahi Chem. Ind. Co. Ltd, Japanese Patent 03120270. 127 Bobyleva, L.I., Kryukov, S.I., Bobylev, B.N., Liakumovich, A.G., Surovtsev, A.A., Karpov, O.P., Akhmedyanova, R.A., and Koneva, S.A. (1992) Yaroslavskij Polit. Institut, SU Patent 1781218. 128 Mais, F.J., Buysch, H.J., MendozaFrohn, C., and Klausener, A. (1993) (Bayer) EP Patent 543249. 129 Kuran, W. and Listos, T. (1994) Macromol. Chem. Phys., 195 (3), 977. 130 Sakai, T., Kihara, N., and Endo, T. (1995) Macromolecules, 28 (13), 4701. 131 Sakai, T., Tsutsumi, Y., and Ema, T. (2008) Green Chem., 10, 337. 132 Sone, M., Sako, T., and Kamisawa, C. (1999) Jpn. Kokai Tokkyo Koho, Japanese Patent 11335372. 133 Song, J., Zhang, Z., Han, B., Hu, S., Li, W., and Xie, Y. (2008) Green Chem., 10, 1337. 134 Darensbourg, D.J. and Holtcamp, M.W. (1996) Coordin. Chem. Rev., 153, 155. 135 Marquis, E.T. and Sanderson, J.R. (1994) (Texaco Chem. Co.) US Patent 5283365. 136 Yano, T., Matsui, H., Koike, T., Ishiguro, H., Fujihara, H., Yoshihara, M., and Maeshima, T. (1997) Chem. Commun., 1129. 137 Yamaguchi, K., Ebitani, K., Yoshida, T., Yoshida, H., and Kaneda, K. (1999) J. Am. Chem. Soc., 121, 4526. 138 Aresta, M., and Dibenedetto, A. (2001) 221st National Meeting, ACS, San Diego, CA, April 1–5, Organic Division, Abstract no. 220. 139 Aresta, M., Dibenedetto, A., Gianfrate, L., and Pastore, C. (2003) J. Mol. Cat. A: General, 204–205, 245.

References 140 Aresta, M., Dibenedetto, A., Gianfrate, L., and Pastore, C. (2003) Appl. Catal. A: General, 255, 5. 141 Li, Y., Zhao, X.Q., and Wang, Y.J. (2005) Appl. Catal. A: General, 279, 205. 142 Zhang, X., Wei, W., and Sun, Y. (2005) Proceedings of ICCDU VIII, Oslo, Norway, June 20–23. 143 Darensbourg, D.J. and Zimmer, M.S. (1999) Macromolecules, 32 (7), 2137. 144 Sugimoto, H. and Inoue, S. (2004) J. Polym. Sci.: A Polym. Chem., 42, 5561. 145 Sugimoto, H., Ohtsuka, H., and Inoue, S. (2004) Stud. Surf. Sci. Catal., 153, 243. 146 Super, M., Berluche, E., Costello, C., and Beckman, E. (1997) Macromolecules, 30 (3), 368. 147 Fujinami, T., Suzuki, T., Kamiya, M., Fukuzawa, S., and Sakai, S. (1985) Chem. Lett., 199. 148 Trost, B.M. and Angle, S.R. (1985) J. Am. Chem. Soc., 107, 6123. 149 Tascedda, P. and Dunach, E. (1997) Heterocycl. Commun., 3, 427. 150 Tascedda, P., Weidmann, M., Dinjus, E., and Dunach, E. (2001) Appl. Organometal. Chem., 15, 41. 151 Welton, T. (1999) Chem. Rev., 99, 2071. 152 Wasserscheid, P. and Keim, W. (2000) Angew. Chem., Int. Ed., 39, 3772. 153 Sheldon, R. (2001) Chem. Commun., 2399. 154 Brown, R.A., Pollet, P., MecKoon, E., Eckert, C.A., Liotta, C.L., and Jessop, P.G. (2001) J. Am. Chem. Soc., 123, 1254. 155 Zhao, D.-B., Wu, M., Kou, Y., and Min, E.-Z. (2002) Catal. Today, 74, 157. 156 Jairton, D., Roberto, F.D.S., and Paulo, A.Z.S. (2002) Chem. Rev., 102, 3667. 157 Peng, J.-J. and Deng, Y.-Q. (2001) New J. Chem., 25, 639. 158 Calò, V., Nacci, A., Monopoli, A., and Fanizzi, A. (2002) Org. Lett., 4, 2561. 159 Kawanami, H., Sasaki, A., Matsui, K., and Ikushima, Y. (2003) Chem. Commun., 896. 160 Li, F.W., Xiao, L.F., Xia, C.G., and Hu, B. (2004) Tetrahedron Lett., 45, 8307. 161 Sun, J.M., Fujita, S.I., and Arai, M. (2004) J. Organomet. Chem., 690, 3490.

162 Kim, H.S., Jelliarko, P., Lee, J.S., Lee, S.Y., Kim, H., Lee, S.D., and Ahn, B.S. (2005) Appl. Catal. A, 48, 48. 163 Sun, J., Zhang, S.J., Cheng, W.G., and Ren, J.Y. (2008) Tetrahedron Lett., 49 (22), 3588. 164 Sun, J., Ren, J., Zhang, S., and Cheng, W. (2009) Tetrahedron Lett., 50 (4), 423. 165 Kawanami, H. and Ikushima, Y. (2000) Chem. Commun., 2089. 166 Mehnert, C.P. (2005) Chem. Eur. J., 11, 50. 167 Hagiwara, H., Sugawara, Y., Isobe, K., Hoshi, T., and Suzuki, T. (2004) Org. Lett., 6, 2325. 168 Abello, S., Medina, F., Rodriguez, X., Cesteros, Y., Salagre, P., Sueiras, J.E., Tichit, D., and Coq, B. (2004) Chem. Commun., 1096. 169 Mehnert, C.P., Cook, R.A., Dispenziere, N.C., and Afeworki, M. (2002) J. Am. Chem. Soc., 124, 12932. 170 Mehnert, C.P., Mozeleski, E.J., and Cook, R.A. (2002) Chem. Commun., 3010. 171 Hagiwara, H., Sugawara, Y., Hoshib, T., and Suzukib, T. (2005) Chem. Commun., 2942. 172 Xie, H.B., Duan, H.F., Li, S.H., and Zhang, S.B. (2005) New J. Chem., 29, 1199. 173 Wang, J.Q., Yue, X.D., Cai, F., and He, L.N. (2007) Cat. Commun., 8, 167. 174 Jacobson, S.E. (1984) Eur. Pat. Appl. EP 118248. 175 Jacobson, S.E. (1984) Eur. Pat. Appl. EP 117147. 176 Aresta, M., Quaranta, E., and Ciccarese, A. (1987) J. Mol. Catal., 41, 355. 177 Aresta, M. and Dibenedetto, A. (2002) J. Mol. Catal., 182–183, 399. 178 Aresta, M., Ciccarese, A., and Quaranta, E. (1985) C1 Mol. Chem., 1, 267. 179 Aresta, M., Quaranta, E., and Tommasi, I. (1994) New J. Chem., 18, 133. 180 Aresta, M., Dibenedetto, A., and Tommasi, I. (2001) Eur. J. Inorg. Chem., (7), 1801. 181 Aresta, M., Dibenedetto, A., and Tommasi, I. (2000) Appl. Organometal. Chem., 14, 799.

209

210

7 Synthesis of Linear and Cyclic Carbonates 182 Sun, J.M., Fujita, S., Bhanage, B.M., and Arai, M. (2004) Catal. Commun., 5, 83. 183 Sun, J.M., Fujita, S., Bhanage, B.M., and Arai, M. (2004) Catal. Today, 383, 93. 184 Sun, J.M., Fujita, S., Zhao, F., Hasegawa, M., and Arai, M. (2005) J. Catal., 230, 398. 185 Eghbali, N. and Li, C.J. (2007) Green Chem., 9 (3) 213. 186 Wang, J.L., Wang, J.Q., He, L.N., Dou, X.Y., and Wu, F. (2008) Green Chem., 10, 1218. 187 Srivastava, R., Srinivas, D., and Ratnasamy, P. (2003) Catal. Lett., 91, 133. 188 Srinivas, D. and Srivastava, R. (2008) US Patent 7365214. 189 Ono, F., Qiao, K., Tomida, D., and Yokoyama, C. (2007) Appl. Catal. A: General, 333, 107. 190 Steimmig, G. and Wittwer, M. (1993) (IG Farben) US Patent 1907891. 191 Malkenus, J.D. (1956) (Jefferson Chem. Co.) US Patent 2766258. 192 Chitwood, H.C. (1957) US Patent 2784201. 193 Venturello, C. and D’Aloisio, R. (1985) Synthesis, 1, 33. 194 Casadei, M.A., Cesa, S., Feroci, M., and Inesi, A. (1999) New J. Chem., 23, 433. 195 Fianz, G., Gallot, Y., Parrod, J., and Rempp, P. (1962) Polym. Sci. Polym. Chem. Ed., 58, 1363. 196 Rokicki, G., Kuran, W., and Pogorzelska-Marciniak, B. (1984) Monatsh. Chem., 115, 205. 197 Rokicki, G. and Kuran, W. (1984) Bull. Chem. Soc. Jpn, 57, 1662. 198 Lee, E.H., Cha, S.W., Dharma, M.M., Choe, Y., Ahn, J.Y., and Park, D.W. (2007) Korean J. Chem. Eng., 24 (3), 547. 199 Renga, J.M. and Pariana-Pillai, R.A. (1982) US Patent 4332729. 200 Pews, R.G. (1974) J. Chem. Soc. Chem. Commun., 4, 119. 201 Aresta, M., Dibenedetto, A., Dileo, C., Tommasi, I., and Amodio, E. (2003) J. Supercrit. Fluids, 25, 177. 202 Sasaki, Y. (1986) Tetrahedron Lett., 27, 1573.

203 Inoue, Y., Ishikawa, J., Taniguchi, M., and Hashimoto, H. (1987) Bull. Chem. Soc. Jpn, 60, 1204. 204 Iritani, K., Yanagihara, N., and Utimoto, K. (1986) J. Org. Chem., 51, 5499. 205 Inoue, Y., Itoh, Y., Yen, I.-F., and Imaizumi, S. (1990) J. Mol. Catal., 60, L1. 206 Dimroth, P. and Pasedach, H. (1961) German Patent 098953. 207 Dimroth, P., Schefczik, E., and Pasedach, H. (1963) German Patent 1145632. 208 Laas, H., Nissen, A., and Nurrenbach, A. (1981) Synthesis, 958. 209 Kim, H.S., Kim, J.W., Kwon, S.C., Shim, S.C., and Kim, T.J. (1997) J. Organomet. Chem., 545–546, 337. 210 Kwon, S.C., Cho, C.S., Shim, S.C., and Kim, T.J. (1999) Bull. Korean Chem. Soc., 20, 103. 211 Gu, Y., Shi, F., and Deng, Y. (2004) J. Org. Chem., 69, 391. 212 Fournier, J., Bruneau, C., and Dixneuf, P.H. (1989) Tetrahedron Lett., 30, 3981. 213 Joumier, J.M., Fournier, J., Bruneau, C., and Dixneuf, P.H. (1991) J. Chem. Soc., Perkin Trans. I, 3271. 214 Bruneau, C. and Dixneuf, P.H. (1992) J. Mol. Catal., 74, 97. 215 Kayaki, H. and Ikarya, T. (2006) Japanese Patent 2006 137733. 216 Kayaki, Y., Yamamoto, M., and Ikariya, T. (2007) J. Org. Chem., 72 (2), 647. 217 Yanlong, G., Feng, S., and Youquan, D. (2004) J. Org. Chem., 69, 391. 218 Tommasi, I. and Sorrentino, F. (2009) Tetrahedron Lett., 50, 104. 219 Tomishige, K., Yasuda, H., Yoshida, Y., Nurunnabi, M., Li, B., and Kunimori, K. (2004) Green Chem., 6, 206. 220 Tomishige, K., Yasuda, H., Yoshida, Y., Nurunnabi, M., Li, B.,and Kunimori, K.(2004) Catal. Lett., 95 (1–2), 45. 221 Huang, S.Y., Liu, S.G., Li, J.P., Zhao, N., Wei, W., and Sun, Y.H. (2007) J. Fuel Chem. Tech., 35 (6), 701. 222 Huang, S.Y., Ma, J., Li, J., Zhao, N., Wei, W., and Sun, Y. (2008) Catal. Commun., 9 (2), 276. 223 Du, Y., He, L.N., and Kong, D.L. (2008) Catal. Commun., 9 (8), 1754.

References 224 Su, W.Y. and Speranza, G.P. (1991) US Patent 5003084. 225 Yutaka, K., Takashi, O., Masahayu, D., Kenichi, K.I., and Atsushi, O. (1993) EP Patent 0581131. 226 Doya, M., Ohkawa, T., Kanbara, Y., and Okmota, A. (1994) US Patent 5349077. 227 Zhao, X.Q., Zhang, Y., and Wang, Y.J. (2004) Ind. Eng. Chem. Res., 15, 4038. 228 Wei, T., Wang, M., Wei, W., Sun, Y., and Zhong, B. (2003) Fuel Proc. Tech., 83 (1–3) 175. 229 Li, Q., Zhao, N., Wei, W., and Sun, Y. (2004) Stud. Surf. Sci. Catal., 153, 573. 230 Ga, Z.W., Wang, S.F., and Xia, C.G. (2009) Chin. Chem. Lett., 20 (2), 131. 231 Hamaguchi, S., Yamamura, H., Hasegawa, J., and Watanabe, K. (1985) Agric. Biol. Chem., 49, 1509. 232 Pfaendler, H. and Muller, F. (1992) Synthesis, 23, 350. 233 Bruson, H. and Riener, T. (1952) J. Am. Chem. Soc., 74, 2100. 234 Bell, J., Currier, V., and Malkemus, J. (1959) US Patent 2915529. 235 Sugita, A. (1994) Japanese Patent 6329663. 236 Claude, S., Zephirin, M., Yoo, J., and Gaset, A. (2000) US Patent 6025504. 237 Okutsu, M. and Kitsuki, T. (2001) EU Pat. Appl EP 1156042. 238 Aresta, M., Dibenedetto, A., Nocito, F., Ferragina, C., and Dubois, J.L. (2008) (Arkema) EU Patent Application 08305653.1-2117. 239 Teles, J., Rieber, N., and Harder, W. (1994) US Patent 5359094. 240 Vieville, C., Yoo, J.W., Pelet, S., and Mouloungui, Z. (1998) Catal. Lett., 56, 245. 241 Aresta, M., Dibenedetto, A., Nocito, F., and Pastore, C. (2006) J. Mol. Catal. A, 257, 149. 242 Sylvain, C., Mouloungui, Z., Yoo, J-W., and Gaset, A. (2001) ) EU Pat. Appl. EP 0955298B1. 243 Munehisa, O. (2007) (Kao Corp., Japan) Jpn. Kokai Tokkyo Koho, Japanese Patent 2007039347. 244 Munehisa, O. and Tomohito, K. (2000) (Kao Corp., Japan) PCT Int. Appl. WO 2000050415. 245 Yoo, J.W. and Moulongui, Z. (2003) Stud. Surf. Sci. Catal., 146, 757.

246 Soga, K., Hosada, S., Tazuke, Y., and Ikeda, S.J. (1976) Polym. Lett. Ed., 14, 161. 247 McMullen, C.H., Nelson, R.J., Ream, C.B., and Sims, J.A. (1982) US Patent 4314945. 248 BockMann, A., Vernaleken, H., Bottenbruch, L., Rudolph, H., and Schnell, H. (1969) DE Patent 1280240. 249 Kem, K. (1981) US Patent 4261922. 250 Nava, H. (1997) US Patent 5679871. 251 Strege, P. (1982) US Patent 4310707. 252 Strege, P. and Doorakian, G. (1982) US Patent 4310708. 253 Dressler, H. (1991) US Patent 5059723. 254 King, S. (1992) US Patent 5104982. 255 Yoshino, T., Inaba, S., Komura, H., and Ishido, Y. (1997) J. Chem. Soc., Perkin Trans. I, 1266. 256 Apicella, B., Serio, M. Di, Fiocca, L., Po, R., and Santacesaria, E. (1998) J. Appl. Polym. Sci., 69, 2423. 257 Angeles, E., Santillan, A., Martinez, I., Ramirez, A., and Moreno, E. (1994) Synth. Commun., 24, 244. 258 Henkelmann, J. and Ruhl, T. (1998) US Patent 5783706. 259 Patil, Y.P., Tambade, P.J., Jagtap, S.R., and Bhanage, B.M. (2008) J. Mol. Catal. A: Chemical, 289 (1–2), 14. 260 Dyer, E. and Scott, H. (1956) J. Am. Chem. Soc., 79, 672. 261 Lee, J. and Litt, M. (2000) Macromolecules, 33, 1618. 262 Pizzi, A. and Stephanou, A. (1993) J. Appl. Polym. Sci., 49, 2157. 263 Stephanou, A. and Pizzi, A. (1993) Holtzforschung, 47, 501. 264 Tohmura, S. and Higuchi, M. (1995) Mokuzai Gakkaishi, 41, 1109. 265 Pizzi, A., Garcia, R., and Wang, S. (1997) J. Appl. Polym. Sci., 66, 255. 266 Park, B. and Riedl, B. (2000) J. Appl. Polym. Sci., 77, 841. 267 Stevenson, J., Machin, J., and Dyke, D. (1983) US Patent 4416694. 268 Knifton, J.F. (1987) US Patent 4661609. 269 Romano, U. and Melis, U. (1997) US Patent 4062884. 270 Green, M.J. (1985) US Patent 4559180. 271 Brunelle, D.J. and Smith, W.E. (1994) US Patent 4349486. 272 French, L.K. and Gilpin, J.A. (1999) US Patent 3642858.

211

212

7 Synthesis of Linear and Cyclic Carbonates 273 Knifton, J.F. (1988) US Patent 4734518. 274 Feng, X., Lu, X., and He, R. (2004) Appl. Catal., 272, 347. 275 Bhanage, B.M., Fujita, S.I., Ikushima, Y., and Arai, M. (2001) Appl. Catal., 219, 259. 276 Bhanage, B.M., Fujita, S.I., Ikushima, Y., Torii, K., and Arai, M. (2003) Green Chem., 5, 71. 277 Zhu, H., Chen, L.B., and Jiang, Y.Y. (1995) Polym. Adv. Technol., 7, 701. 278 Wei, T., Wang, M., Wei, W., Sun, Y., and Zhong, B. (2003) Green Chem., 5, 343. 279 Wang, H., Wei, W., Wang, X., and Sun, Y. (2004) Petrochem. Technol., 33, 512. 280 Besson, M. and Gallezot, P. (2003) Catal. Today, 81, 54713. 281 Wanga, H., Wanga, M., Zhao, N., Wei, W., and Sun, Y. (2005) Catal. Lett., 105, 3. 282 Wei, T., Wang, M., Wei, W., Sun, Y., and Zhong, B. (2003) Fuel. Proc. Technol., 83, 175. 283 Li, Y., Zhao, X., and Wang, Y. (2005) Appl. Catal., 279, 205. 284 Chang, Y., Jiang, T., Liu, Z., Wu, W., Gao, L., Li, J., Gao, H., Zhao, G., and Huang, J. (2004) Appl. Catal., 263, 179.

285 Cui, H., Wang, T., Wang, F., Gu, C., Wang, P., and Dai, Y. (2003) Ind. Eng. Chem. Res., 42, 3865. 286 Cui, H., Wang, T., Wang, F., Gu, C., Wang, P., and Dai, Y. (2004) J. Supercrit. Fluids, 30, 63. 287 Watanabe, Y. and Tatsumi, T. (1998) Micropor. Mesopor. Mater., 22, 399. 288 Aresta, M., Dibenedetto, A., and Pastore, C. (2004) Stud. Surf. Sci. Catal., 153, 221. 289 Carothers, W.H. and Van Natta, F.J. (1930) J. Am. Chem. Soc., 52, 314. 290 Albertsson, A.C. and Sjoeling, M. (1992) J. Macromol. Sci. Pure Appl. Chem., A29, 43. 291 Sarel, S. and Pohoryles, L.A. (1958) J. Am. Chem. Soc., 80, 4596. 292 Sarel, S., Pohoryles, L.A., and BenShoshan, R. (1959) J. Org. Chem., 24, 1873. 293 Clements, J., Klein, H., and Marquis, T. (2003) Huntsman Corporation US Patent 011330. 294 Rokicki, G., and Kowalczyk, T., and Glinski, M. (2000) J. Polym., 32 (5), 381. 295 Ohno, M. and Yamagiwa, S. (2000) US Patent 6054596.

213

8 Polymers from Carbon Dioxide: Polycarbonates, Polythiocarbonates, and Polyurethanes Donald J. Darensbourg, Jeremy R. Andreatta, and Adriana I. Moncada

8.1 Introduction

Prominent among the processes that represent technologically viable uses of carbon dioxide (CO2), both as a solvent and a monomer, is the production of biodegradable copolymers. Although CO2 is thermodynamically and kinetically very stable, it does undergo reactions with highly reactive molecules, most notably among which are transformations with ring-strained heterocycles such as propylene oxide (PO) and cyclohexene oxide (CHO) (Scheme 8.1). As noted in Scheme 8.1, these processes are generally accompanied by the production of differing amounts of ﬁve-membered cyclic carbonates, with aliphatic epoxides affording larger quantities than alicyclic epoxides. Of importance, both the copolymer and its corresponding cyclic carbonate afforded from the alternating coupling of epoxides and CO2 have industrial applications. That is, polycarbonates are known for their outstanding properties, including strength, lightness, durability, high transparency, heat resistance, and good electrical insulation [1]. Cyclic carbonates ﬁnd uses as high-boiling solvents and reactive intermediates [2]. An alternative source of polycarbonate derived from CO2 involves the alternating copolymerization of oxetane and CO2 (Equation 8.1). O O

+ CO2

oxetane

catalyst

O

O O C

n

Poly(trimethylene carbonate)

+

O

O

TMC (trimethylene carbonate) (8.1)

The ring-strain energy of oxetane is less than that of PO (106.7 kJ mol−1 versus 114.2 kJ mol−1); hence, its copolymerization with CO2 is less favored thermodynamically [3]. Nevertheless, the copolymerization of oxetane and CO2 occurs readily under similar catalytic conditions, producing poly(trimethylene carbonate), Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

214

8 Polymers from Carbon Dioxide

Scheme 8.1 Coupling reactions of CO2 and epoxides to afford polycarbonate and cyclic products.

an important component of biodegradable thermoplastics elastomers which have numerous clinical applications [4]. A signiﬁcant difference between this coupling reaction and the epoxide/CO2 process is that, in this instance, the six-membered cyclic carbonate is thermodynamically unstable relative to the copolymer. In a closely related process, the synthesis of polytrithiocarbonate by the alternating copolymerization of episulﬁde and carbon disulﬁde has provided a novel source of sulfur-enriched polymers (Equation 8.2). S

S S + CS2

catalyst

S

S

+

S (8.2)

S

n

Because of their high refractive index, these materials possess desirable optical and thermal properties, in addition to being impervious to a variety of chemicals [5]. Analogous reactions involving carbon disulﬁde and epoxides have provided extremely interesting examples of atom-exchange polymerization processes. The ﬁnal topic to be covered in this chapter will be the production of poly(urethane-amine)s from CO2 and aziridine derivatives (Equation 8.3). H

O

N + CO2

H N

C N H

O

n

(8.3) 1-n

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

8.2 Historical Perspective

The ﬁrst report of the copolymerization of an epoxide, namely, ethylene oxide and CO2 is contained in a patent by Stevens [6]. However, this process, when carried out in the presence of polyhydric phenols, provided polymers which were viscous liquids or waxes possessing copious polyether linkages with only a few incorporated CO2 units. The earliest metal-catalyzed copolymerization of epoxides and CO2 was reported in 1969 by Inoue and coworkers, who employed a heterogeneous catalyst system derived from a 1 : 1 mixture of diethylzinc and H2O [7, 8]. Subsequently, Kuran and coworkers investigated a group of related catalysts prepared from diethylzinc and di- and triprotic sources such as pyrogallol, with a slight improvement over Inoue’s system for the production of poly(propylene carbonate) from PO and CO2 [9]. Since these early studies, a variety of other heterogeneous catalyst, have been reported to promote the copolymerization of PO and CO2, with the most active of these being air-stable zinc derivatives of dicarboxylic acid [10, 11]. These catalysts are readily synthesized by the reaction of zinc hydroxide or zinc oxide with dicarboxylic acids in toluene. Recently, the crystal structure of zinc glutarate has been determined from its powder pattern of polycrystalline material [12] or single crystals [13], and shown to be a layered structure of zinc ions with bridging dicarboxylates between the layers (see Figure 8.1). This latter catalyst system is patented [14], and because of its robust nature and inexpensive starting reagents, it is used commercially for the production of poly(ethylene carbonate) and poly(propylene carbonate) from ethylene oxide or PO and CO2 [15]. Unfortunately, this heterogeneous catalyst system requires very high catalyst loadings, and generally provides an extremely poor control of molecular weight, with polydispersities greater than 3 or 4. Of importance to environmental concerns, this catalyst system performs equally well in supercritical carbon dioxide (scCO2) or halogenated organic solvents [16]. The scope of this chapter will be to examine discrete, well-characterized metalbased catalysts for the incorporation of CO2 as a monomer in the preparation of polymeric materials. The coupling of CO2 and epoxides has recently been the subject of several comprehensive reviews, which will not be reiterated herein [17–19]. Instead, attention will be focused on some of the more recent major contributions to this area over the past decade.

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

Before examining the more recent developments involving single-site metal catalysts for the copolymerization of CO2 and epoxides, it may be worthwhile discussing some general features of these processes, and also deﬁning some commonly used terms. First of all, the reaction sequences depicted in Scheme 8.1 are not

215

216

8 Polymers from Carbon Dioxide

Figure 8.1 Crystal structure of zinc glutarate prepared from ZnO and glutaric acid in toluene. Structure determined from X-ray powder pattern (view along the c-axis) [12].

completely correct. That is, these coupling reactions often lead to copolymers with varying quantities of randomly distributed ether linkages (Equation 8.4). O O

R O R

+ CO2

catalyst

O O

O

n

R

+ O m

O R (8.4)

Most often, the extent of completely alternating copolymer formation, expressed as 100% CO2 linkages or 50% CO2 content, is very high. With regards to the selectivity of the coupling reaction for copolymer versus cyclic carbonate production, two observations are consistently found, regardless of the catalyst. First, aliphatic epoxides are more prone to cyclic carbonate formation than alicyclic epoxides; for example, PO affords propylene carbonate more readily than CHO provides cyclohexene carbonate. Second, in either instance, since it has been shown that the activation barriers for cyclic carbonate production are higher

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

than that of copolymer formation, an increase in temperature will lead to an increase in cyclic carbonate production. Recall that the thermodynamically most stable product from the coupling of CO2 and epoxides is the ﬁve-membered cyclic carbonate. Relevant to this issue, it is worthwhile noting here that only a limited number of epoxides are known to selectively afford polycarbonates from reaction with CO2. These include:

The alicyclic epoxide, limonene oxide, which is obtained from a renewable resource has shown modest activity compared to CHO for reaction with CO2 to provide a copolymer. This signiﬁcant decrease in reactivity is presumably due to the steric inﬂuence of a disubstitution at one of the ipso carbon centers. Of course, in either highly selective reaction, where complete formation of copolymer or cyclic carbonate occurs, the process displays 100% atom economy. The environmental attractiveness of this process is further enhanced by the fact that reactions are generally carried out in the absence of an organic cosolvent, that is, in CO2-swollen epoxide solutions. Finally, it is beneﬁcial at this point to deﬁne two terms which will be widely alluded to in future discussions, namely, turnover frequency (TOF) and polydispersity. TOF is the moles of epoxide consumed per mole catalyst per hour. Because these copolymerization processes occur on a rather slow timescale, TOFs are usually expressed in units of h−1. Polydispersity is also referred to as a molecular weight – – – distribution and is deﬁned as Mw / Mn, where Mw is the weight average molecular 2 – weight ⎛⎜ ΣNiMi ⎞⎟ and Mn is the number average molecular weight ⎛⎜ ΣNiMi ⎞⎟ . Ni ⎝ ΣNi ⎠ ⎝ ΣNiMi ⎠ equals the number of chains with mass Mi, and Mi equals the mass of the chain. Mw is always >Mn. The ﬁrst single-site metal catalyst which was shown to homogeneously catalyze the coupling of epoxides and CO2 was (tpp)AlCl (tpp = tetraphenylporphyrin) in the presence of a quaternary organic salt or triphenylphosphine [20]. Although this catalytic system was extremely slow at ambient temperature, copolymers from ethylene oxide, PO, and CHO and CO2 were obtained that possessed very narrow molecular weight distributions (polydispersity = 1.06–1.14). The low reactivity of

217

218

8 Polymers from Carbon Dioxide

these aluminum porphyrin derivatives has been utilized by Chisholm and Zhou to perform comprehensive studies of the mechanistic aspects of this process [21]. As will be discussed later, the mechanistic aspects of the CO2/epoxide copolymerization reaction catalyzed by Schiff base and porphyrin metal complexes have much in common. Analogous chromium(III) porphyrin complexes exhibit a greater activity for copolymerizing CHO and CO2 [22]. Indeed, a ﬂuorinated (tpp)CrCl was found to be soluble in sc CO2, thereby affording a polycarbonate with a very high percentage of carbonate linkages [23]. A revitalization of interest in the copolymerization of epoxides and CO2 can be traced to studies involving discrete zinc complexes. Several well-deﬁned zinc monomeric and dimeric derivatives have been shown to be effective catalysts for the coupling of CHO and CO2 to afford copolymers with high degrees of CO2 incorporation. These include reports of sterically encumbering bis-phenoxide derivatives of zinc [24], a highly ﬂuorinated zinc carboxylate [25], and zinc βdiiminate complexes [26] (see Figure 8.2). Of these zinc complexes, the β-diiminate derivatives exhibited the greatest catalytic activity towards poly(cyclohexene carbonate) production. Nevertheless, their ability to catalyze copolymer formation was quite sensitive to variations in the electronic and steric characteristics of the ligand. For example, the derivative where

Figure 8.2 (a) Bis-2,6-diisopropylphenoxide zinc (THF)2; (b) Highly ﬂuorinated zinc carboxylate; (c) β-Diiminate zinc complex.

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

Figure 8.3 Transition state of the epoxide ring-opening step (P = growing polymer chain) for: (a) dizinc catalyst; (b) dimagnesium catalyst.

R1 = CN and R2 and R3 = Me and iPr, respectively, displayed a TOF of 2290 h−1 at 323 K, whereas when R1 = H and R2 = R3 = Et, a corresponding TOF of 239 h−1 was observed. Similarly, subtle changes in the nature of the β-diiminate ligand has led to efﬁcient catalysis of other epoxides, that is, PO [27] and limonene oxide [28], with CO2 to selectively afford polycarbonates. Coates and coworkers have carried out kinetic studies of the alternating copolymerization of CHO and CO2 catalyzed by several of the β-diiminate zinc derivatives [29]. These authors have proposed a bimetallic mechanism to be operative, which is consistent with their experimental observations, including the large differences in activity noted for a series of structurally closely related catalysts. It was proposed that one zinc center would coordinate and activate the epoxide substrate, while the second zinc center would provide the propagating carbonate species to ring-open the epoxide. This proposal is represented by the transition state depicted in Figure 8.3a. In a similar manner, rate studies of the copolymerization of CHO and CO2 catalyzed by dinuclear magnesium complexes have suggested the involvement of the two magnesium centers acting in a synergistic manner to promote the reaction, as shown in Figure 8.3b [30]. Much of the present understanding of the mechanistic aspects of the copolymerization process has been gained from kinetic studies utilizing in situ infrared (IR) spectroscopy. Figure 8.4 illustrates a view of one such system used for monitoring these processes, namely the ASI® ReactIR 1000 reaction system with a stainless-steel Parr® autoclave. This reaction system (and its newer versions which have been modiﬁed for high-pressure applications) allows for the collection of IR spectra in the mid-IR range, without disruptive sampling. Figure 8.5 depicts the type of data collected for growth of the carbonate band of poly(cyclohexene carbonate) for a typical copolymerization reaction catalyzed by a (salen)CrCl catalyst at 5.5 MPa CO2 pressure and 353 K. A detailed description of how this in situ attenuated total reﬂectance-Fourier transform infrared (ATR-FTIR) spectroscopy system operates, together with its inherent limitations, can be found elsewhere [31].

219

220

8 Polymers from Carbon Dioxide

Figure 8.4 The ASI® ReactIR 1000 reaction analysis system with a stainless-steel Parr® autoclave modiﬁed with permanently mounted ATR crystal (SiComp) at the bottom of the reactor (purchased from Mettler Toledo).

ADS 0.2500 0.2000 0.1500 0.1000 0.05000 0.0 1850

1800

1750 1700 Wavenumber (cm–1)

6.0 5.0 4.0 3.0 2.0 1.0 0.0 16.50 Hours

Figure 8.5 Representative three-dimensional stack plot of the infrared spectra collected every 3 min during the reaction of CO2 and TMSO (353 K, 5.5 MPa pressure).

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

Figure 8.6 (a) General structure of (salen)CrX catalyst; (b) X-ray deﬁned structure where R3 = R4 = tBu, and R1 = R2 = R,R-cyclohexylene.

Figure 8.7 Skeletal representation and crystallographic structure of the (tmtaa)CrCl catalyst for the copolymerization of epoxides and CO2.

Most recent studies of the copolymerization of epoxides and CO2 have centered on processes catalyzed by (salen)MX (M = Cr, Co, Al) in the presence of quaternary organic salts. Figure 8.6 depicts a generic structure of the chromium complex, where the accompanying quaternary organic salts are generally n-Bu4N+X− or PPN+X− (PPN+ = Ph3P=N+=PPh3). In the absence of sterically encumbering groups on the diimine backbone (R1 and R2), little difference in reactivity was noted; that is, the R,R-cyclohexylene, ethylene, and phenylene backbones exhibited similar catalytic activity [32]. On the other hand, substituents on the 3,5-positions of the phenolic ligands (R3 and R4) signiﬁcantly affected the reaction rates, with electron-donating substituents enhancing catalytic activity. Furthermore, very recent studies have shown the saturated version of the salen ligand, salan (N,N’-disubstituted bis(aminophenoxide)), to be even more effective as a catalyst for the copolymerization of epoxides and CO2, especially for PO [33]. In addition, the tetramethyltetra-annulene ligand, which is signiﬁcantly more electron-donating and less expensive than its porphyrin analogues, when bound to chromium(III) (Figure 8.7), has been shown to be more effective than salen derivatives at catalyzing the copolymerization of CHO and CO2 [34]. The generally accepted mechanism for the copolymerization of epoxides and CO2 catalyzed via these tetradentate, approximately planar N2O2−2 or N4−2-ligated metal complexes in the presence of anionic initiators is outlined in Scheme 8.2.

221

222

8 Polymers from Carbon Dioxide

X

X

Cr

Cr

X epoxide

Nuc or

Cr

Cr

Nuc Nuc

Nuc

X(Nuc) Cr

CO2

O epoxide

C

O

X

X

Nuc

Cr

Cr

O

O

Nuc

X

O

O

O Nuc(X)

CO 2

X(Nuc)

= Salen Ligand

Cr O P Scheme 8.2 Mechanistic aspects of cyclohexene oxide/CO2 copolymerization in the presence of (salen)CrX and cocatalyst (Nuc = anion).

Although neutral nucleophiles (e.g., DMAP and Cy3P) have been shown to be effective as cocatalysts for these processes, in these instances there are initiation periods prior to achieving the maximum rate of copolymer production. This is best illustrated in Figure 8.8 for the formation of the copolymer from CHO and CO2, as revealed by IR traces during in-situ IR monitoring. The TOFs at 353 K for the copolymerization of CHO and CO2 as a function of the initiating anion in the presence of (salen)CrX (R1 = R2 = H and R3 = R4 = tBu) are listed in Table 8.1. These data show that the catalytic activity decreases in the order N3 > Cl > Br > I. It is worthy of note that TOFs as reported here and elsewhere in the literature are highly dependent on the time period of the measurement. That is, the highest TOFs will be obtained during the initial period of high polymer production. As seen in Table 8.1, the copolymerization process is dependent on the nature of both the anionic initiator and the salen ligand. For the copolymerization of CHO and CO2, the catalytic activity has been optimized at a TOF of 1153 h−1 at 353 K and 3.5 MPa of CO2 for the (salen)CrX complex, where X = N3 and R1, R2 = (1R,2R)– C4H8–, R3 = OCH3, and R4 = tBu in the presence of PPNN3 [35].

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

Figure 8.8 Comparison of in situ infrared proﬁles of copolymer production from CO2 and cyclohexene oxide utilizing the catalyst in Figure 8.6b and the three classes of cocatalysts. 䉲 = PPN+Cl−; 䊊 = PCy3; 䊉 = N-MeIm. Catalytic activity for the copolymerization of cyclohexene oxide and CO2 in the presence of one equivalent of PPNX cocatalyst.a

Table 8.1

X

Turnover frequencyb

N3 Cl Br I OAc HCO3

608 494 420 360 350 280

a)

The carbonate content of all copolymers isolated was greater than 99%. b) Moles epoxide consumed mol−1 catalyst h−1.

An aspect of the mechanism of this copolymerization process which, up to this point is unresolved, results from the general observation that the molecular weights of the copolymers, as determined for the puriﬁed copolymers, are signiﬁcantly less than anticipated from a theoretical basis. This discrepancy is generally assumed to be due to a chain-transfer process with trace quantities of water (Equation 8.5). X

X

M

+ H2O

O

M OH

P

new initiator

+ HO

P (8.5)

223

224

8 Polymers from Carbon Dioxide

X X X

O

O

Cr

epoxide

O

Cr

CO2

O C

O

O

C

O

Cr O

C

O

O

O

O

Nuc

Nuc

Nuc

Scheme 8.3

C O

Simultaneous chain propagation from both sides of the (salen)Cr plane.

Indeed, there is evidence that when extreme care is taken to exclude any trace water impurities, the molecular weights of the copolymers produced are increased – that is, they approach the expected weights [36]. Alternatively, it has been suggested that this difference in copolymer molecular weights is possibly the result of chain propagation on both sides of the approximately planar salen ligand (see Scheme 8.3) [37]. Based on the fact that the rate of copolymerization is heavily dependent on the nature of the X group attached to the metal center, this conclusion seems unlikely. Nevertheless, an experiment has recently been designed which utilizes the closely related (tmtaa)CrCl catalyst system to address this issue directly [38]. In order to accomplish this, two electronically identical tetramethyltetraazaannulene complexes were synthesized, one containing a ligand with a protected underside (Figure 8.9a) and another with a ﬂexible underside (Figure 8.9b). Although the (tmtaa)CrCl catalyst derived from the ligand in Figure 8.9a has been shown to readily bind small anions such as chloride, it should not allow chain propagation to occur at this highly sterically restricted side. On the other hand, the (tmtaa)CrCl derivative prepared from the ligand in Figure 8.9b readily binds to large anions on both sides of the tetraaza-annulene plane, indicating a lack of steric hindrance. The catalytic activities for the copolymerization of CHO and CO2 performed under identical conditions utilizing the two closely related (tmtaa)CrCl complexes were found to be indistinguishable, as were the copolymer’s molecular weights. Hence, this is taken to be strong evidence for the lack of dual catalysis behavior at a single metal site, as depicted in Scheme 8.3. Because of the structural differences in the (tmtaa)CrCl and (salen)CrCl complexes, it will be necessary to prepare analogous underside-protected and -unprotected salen complexes, and to repeat these copolymerization studies before reaching a conclusion on these catalyst systems. Relevant to this copolymerization process, upon reacting (salen)CrX complexes in the presence of quaternary organic salts (n-Bu4NX or PPNX), several of the

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

Figure 8.9 (a) X-ray crystal structures of the strapped tmtaa ligand, showing both side-on and end-on perspectives. Thermal ellipsoids are shown at the 50% probability level with selected hydrogens and solvent molecules omitted for clarity; (b) X-ray crystal structure

of strap mimic ligand, illustrating the ﬂexibility of the anisoyl arms to expose the underside of the complex. Thermal ellipsoids are shown at the 50% probability level, with hydrogens omitted for clarity.

isolated anionic metal derivatives have been well-characterized using X-ray crystallography (Figure 8.10) [39]. In addition, an X-ray crystallographic structure of the distorted octahedrally coordinated Na[Co(salen)(N3)2] complex has recently been reported [40]. As noted earlier in the chapter, studies of the copolymerization of CHO and derivatives thereof selectively produced almost completely alternating copolymers at elevated temperatures (e.g., 333–373 K). In order to avoid mostly cyclic carbonate products for reactions involving aliphatic epoxides, low reaction temperatures are required [41–43]. Until very recently, (salen)CoX complexes in the presence of cocatalysts have been most effective for selectively affording copolymers from PO and CO2. Importantly, at ambient temperature these cobalt(III) catalysts have been shown to produce copolymers from PO and CO2 with both high activity and high regioselectivity and stereoselectivity. For example, the cobalt analogue of the chromium complex depicted in Figure 8.6b, in the presence of anionic cocatalysts or

225

226

8 Polymers from Carbon Dioxide

Figure 8.10 Thermal ellipsoid representations of (salen)CrX 2− anions, where the salen ligand contains –OMe and –t-Bu substituents in the 3,5-positions of the phenolates, respectively, with a phenylene diamino-backbone. (a) n-Bu4N+ salt, X = Cl; (b) n-Bu4N+ salt, X = N3.

N,N-dimethylaminoquinoline (DMAQ), was found to afford poly(propylene carbonate) with high regioregularity (head-to-tail connectivity), high molecular weights with narrow polydispersities, and tacticity control (isotactic or syndiotactic) [36, 44–48]. (A detailed description of regioregular and stereoregular copolymers from PO and CO2 is provided later in the chapter.) In addition, the chiral (salen) CoX version of the complex binary catalyst systems (as shown in Figure 8.6b) were seen to react preferentially with (S)-PO over (R)-PO, with krel = 2.8–5.4. A related cobalt porphyrin complex in the presence of the Lewis base cocatalyst, N,N-dimethylaminopyridine (DMAP), has also been shown to exhibit good catalytic activity and selectivity for copolymer formation, from both CHO or PO and CO2, under very mild reaction conditions (ambient temperature and 0.1 MPa of CO2) [49]. Several novel modiﬁed salen derivatives of cobalt(III) have provided convincing evidence for the importance of the propagating copolymer chain staying in the vicinity of the metal center, so as to avoid the formation of cyclic carbonates; this procedure is especially relevant to processes involving the PO monomer. Both, computational and experimental studies have shown that cyclic carbonate formation is enhanced relative to monomer enchainment under conditions where the growing polymer chain is outside the inﬂuence of the metal catalyst [50, 51]. To circumvent this issue, Nozaki and coworkers prepared a salen complex containing a piperidinium end-capping arm (Scheme 8.4) [52]. During the copolymerization of PO and CO2, it was possible to control the formation of the cyclic carbonate, presumably by protonation of the anionic polymer chain when it became dissociated from the metal center. As a consequence, the reaction could be carried out at an elevated temperature (333 K), without the extensive production of propylene carbonate (∼10%). In closely related studies, Lee and coworkers obtained even more dramatic results in attempts to alleviate the generally observed selectivity for the formation of cyclic carbonates from aliphatic epoxides and CO2 [53, 54]. This was achieved by adding cationic charge to the salen ligands of cobalt(III) complexes, as illustrated in Figure 8.11.

8.3 Metal Catalysts for the Copolymerization of Epoxides and CO2

Scheme 8.4 (Salen)Co(III) complex bearing piperidinyl/piperidinium arms for propylene oxide/CO2 copolymerization process.

Figure 8.11 Designer catalysts for propylene oxide and CO2 copolymerization at elevated temperatures.

By using complexes which contained internal anions that served as initiators, Lee and coworkers were able to produce high-molecular-weight copolymers (Mn up to 296 000) with narrow molecular weight distributions (PDI = 1.19), and with a greater than 97% selectivity for copolymer formation. Furthermore, this methodology has provided a strategy for catalyst separation and recovery (vide infra). A recently reported alternative system for the effective copolymerization of PO and CO2 involved the use of chiral (salan)CrX catalysts in the presence of quaternary organic salts. Using this catalyst system, which represents a saturated version of the corresponding Schiff-base salen ligand shown in Figure 8.6, Lu and coworkers have copolymerized racemic PO and CO2 under mild reaction conditions with ∼95% head-to-tail linkages and modest enantioselectivity; that is, with enantiomeric excess (ee) values ranging from 22.5% to 70.9% (krel = 1.6–7.1) [33]. In general, the selectivity for copolymer versus cyclic carbonate formation at 298– 323 K was greater than 90%. In addition, an achiral (salan)CrX catalyst with an ethylene diamine backbone has been similarly shown to afford a PO/CO2 copolymer with greater than 95% head-to-tail regioselectivity, as well as being very effective for providing diblock copolymers between poly(propylene carbonate) and poly(cyclohexylene carbonate) [55]. In this latter instance, the chromium(III) salan complex displayed a cis geometry upon the addition of a sixth ligand, which probably accounts in part for the difference in behavior of (salen)CrX and (salan)CrX catalyst systems (Figure 8.12).

227

228

8 Polymers from Carbon Dioxide

t-Bu

t-Bu N O

N Cr t-Bu

O

Cl

t-Bu Figure 8.12 Proposed structure of (salan)CrCl, where 䊐 indicates the site for epoxide binding [55].

8.4 Metal Catalysts for the Copolymerization of Oxetanes and CO2 8.4.1 Early Studies

The origin of the metal-catalyzed copolymerization of oxetane and CO2 can be traced back to the early efforts of Professors Koinuma and Hirai at the University of Tokyo, who were inspired by their earlier studies on the copolymerization of oxiranes (epoxides) and CO2 catalyzed by organoaluminum catalysts to yield polycarbonates [56]. Hence, Koinuma and Hirai thought that it might be possible to generate aliphatic polycarbonates not only from three-membered cyclic ethers, but also from four-membered cyclic ethers, oxetanes. Initially, poly(trimethylene carbonate) (poly(TMC)) was prepared by employing a ternary catalyst system composed of triethylaluminum, water, and acetylacetone (2 : 1 : 1) [57], but unfortunately the system was plagued by poor yields of copolymers and a low degree of CO2 incorporation. Consequently, an anionic coordination mechanism was proposed for the formation of poly(TMC). Subsequently, Baba, at Osaka University, reported the use of organotin halides and Lewis bases as catalytic systems, producing lowmolecular-weight polycarbonates from the coupling of oxetane and CO2 [58]. Baba also used tetraphenylstibonium iodide for the selective synthesis of TMC from oxetane and CO2 [59]. A few years later, Baba reported a more comprehensive description of the coupling reaction of oxetane and CO2 catalyzed by organotin iodides and Lewis bases as catalysts [60]. It was shown in this instance, that the choice of Lewis base which coordinated to the organotin iodides affected the catalytic activity and selectivity, that is, poly(TMC) and/or trimethylene carbonate. Whilst complexes with Bu3P yielded polycarbonate, the combination of Bu3SnI with Bu3P=O yielded TMC exclusively in good yields. A reaction mechanism, as proposed by Baba and coworkers, is illustrated in Scheme 8.5.

8.4 Metal Catalysts for the Copolymerization of Oxetanes and CO2

Scheme 8.5 Reaction mechanism for production of poly(TMC) from oxetane/CO2 using a Bu3SnI catalyst in the presence of Lewis bases.

Thus, in the ﬁrst step the oxetane is ring-opened by an organotin iodide complex to produce an organotin iodopropoxide(II) intermediate, which then undergoes insertion of CO2 into the Sn–O bond, to generate an organotin carbonate adduct. The TMC is produced by a back-biting mechanism, and the organotin iodide complex is regenerated. The coordination of phosphine or phosphine oxide ligands to organotin iodide compounds activates the Sn–I bond by enhancing the nucleophilicity of the halide initiator, allowing the oxetane ring to be opened. Furthermore, the polymerization of preformed TMC is proposed to occur by free organotin iodide complexes in solution. This hypothesis was supported by the fact that, in the presence of a large excess of Bu3P=O, its dissociation from Bu3SnI was suppressed and no polymerization occurred. In contrast, complexes of Bu3P and Bu3SnI were found to be too unstable to suppress polymerization, even in the presence of a large excess of phosphine. 8.4.2 Recent Studies using Transition Metal Catalysts

Because the ring-strain energy of PO and oxetane do not differ greatly (114.2 versus 106.7 kJ mol−1, respectively), it would be expected that (salen)CrX complexes in the presence of quaternary organic salts would serve as catalysts for the copolymerization of oxetanes and CO2. This transformation is of particular interest, since in this case the cyclic carbonate byproduct, TMC, unlike the ﬁve-membered cylic carbonates, can be ring-opened and transformed into the same alternating copolymer. In 2006, efforts were commenced to examine the efﬁciency of metal– salen complexes for the copolymerization of oxetane and CO2. Initially, the catalytic activity of the (salen)M(III)Cl derivatives (Figure 8.13) towards the copolymerization of oxetane and CO2 in the presence of n-Bu4NCl as cocatalyst was investigated.

229

230

8 Polymers from Carbon Dioxide

N

N M

O

t-Bu

Cl

O

t-Bu

t-Bu

t-Bu M = Cr, Al

Figure 8.13 General structure of the metal(III) salen catalysts employed in the copolymerization of oxetanes and CO2.

Cr

-

Cl

Cl +

N3

-

2009 cm-1

N3 N3

Cr N3

-

Cr

+

Cl -

N3 -1

2051 cm

2057 (sh), 2047 cm-1

Scheme 8.6 Infrared frequencies in the νN3 stretching region for various species present in the reaction of (salen)CrCl and azide ions.

These reactions were carried out at 383 K, using a CO2 pressure of 3.5 MPa and 2 equiv. n-Bu4NCl. The (salen)Cr(III)Cl complex (see Figure 8.13) was shown to be more active (TOF = 41.2 h−1) than its aluminum salen analogue (TOF = 8.59 h−1) when catalyzing this coupling reaction [61]. Further optimization of the catalytic activity of the Cr(III) system was achieved by utilizing a salen ligand with tert-butyl groups in the 3,5-positions of the phenolate rings and a cyclohexylene backbone for the diimine, along with an azide ion initiator. In this manner, a selectivity of 100% towards poly(TMC) formation was obtained after 24 h, with a TOF OF 35 h−1 and 96% CO2 incorporation, for a reaction performed at 383 K and 3.5 MPa of CO2 pressure [62]. It is important to note here that minimal quantities of ether linkages were observed in the resulting copolymers as a result of the consecutive oxetane ring-opening processes. Fundamental to a better understanding of the mechanism of the coupling reaction of oxetane and CO2 was an investigation into the initiation step of the process. As mentioned earlier, upon dissolution of a typical ﬁve-coordinate (salen)CrCl complex in the presence of an azide-based cocatalyst in a weakly interacting solvent, the formation of a six-coordinate complex of the type [trans-(salen)Cr(N3)2] [PPN or n-Bu4N] was evident. This reaction was monitored using solution IR spectroscopy in the νN3 stretching region (Scheme 8.6). In addition, the X-ray crystal structure of the complex resulting from the addition of n-Bu4NN3 to (salen)CrCl has been described (see Figure 8.10) [39]. Upon adding an excess of the cyclic ether monomer to the six-coordinate anionic complex

8.4 Metal Catalysts for the Copolymerization of Oxetanes and CO2

(salen)CrX

+

X-

(salen)CrX2monomer

(salen)CrX(monomer)

+

X-

Scheme 8.7 Reaction of cyclic ether monomers with trans-(salen) CrX 2− derivatives.

N3

O

N3

N3

Cr

Cr

N3

O

+

N3

110 °C

Cr O

2057(sh) 2047 cm-1 2061 cm-1

N3

2100 cm-1

Scheme 8.8 Formation of oxetane adduct with (salen)CrN3 and subsequent ring-opening process.

of the type [trans-(salen)CrX2][PPN or n-Bu4N], an equilibrium process was established, as shown in Scheme 8.7. The equilibrium position and subsequent rate of ring-opening differed signiﬁcantly with the nature of the cyclic ether monomer. For example, the monomers, PO and oxetane, which have similar steric requirements and signiﬁcantly different pKb values (6.94 and 3.13, respectively) [63], would be expected to bind to the metal center quite differently. As shown in Scheme 8.8, solution IR spectroscopy revealed − that, in the presence of a large excess of oxetane monomer, the (salen)Cr ( N3 )2 species existed as an equilibrium mixture, greatly favoring the neutral oxetanebound species [39]. Pertinent to this step, a (salen)CrCl complex with an oxetane molecule bound axially to the chromium center was successfully characterized by X-ray crystallography (Figure 8.14) [62]. Furthermore, the azide ring-opening of oxetane could only be detected at high temperatures (383 K) by solution IR spectroscopy (organic azide = 2100 cm−1) of (salen)Cr(III)(N3)·oxetane species in tetrachloroethane solution. By contrast, CHO and PO monomers were shown to undergo ringopening at ambient temperature, with PO being ring-opened by the azide-based catalyst system at a signiﬁcantly faster rate than the corresponding process involving CHO. The mechanistic aspects of the catalytic coupling of oxetane and CO2 have been investigated by monitoring the process by in situ IR spectroscopy, by observing the growth of the copolymer’s νC=O band at 1750 cm−1, and that of TMC’s νC=O band at 1770 cm−1. The catalytic system employed in these studies was a (salen)CrCl along with n-Bu4NN3 as the initiator. The reactions were carried out in toluene at

231

232

8 Polymers from Carbon Dioxide

Figure 8.14 Thermal ellipsoid plot of (salen)CrCl·oxetane, where the salen ligand contains -t-Bu substituents in the 3,5 positions of the phenolates, respectively, with a phenylene diamino backbone.

38 K and 35 MPa CO2 pressure. Figure 8.15 illustrates a typical reaction proﬁle of the growth of the copolymer’s νC=O band at 1750 cm−1 as a function of time. Importantly, an IR band (νC=O 1770 cm−1) due to TMC was detected during the early stages of the reaction, but this completely vanished after a few hours. This was also observed by intermittent sampling for 1H nuclear magnetic resonance (NMR) analysis. From temperature-dependent studies, ΔH≠, ΔS≠, and ΔG≠ values of 45.6 kJ mol−1, 161.9 J mol−1 deg−1, and 106.7 kJ mol−1, respectively, were obtained. A kinetic study of the ring-opening polymerization (ROP) of trimethylene carbonate similarly afforded ΔG≠ at 383 K of 101.9 kJ mol−1, a value which was very close in energy to that found for the production of poly(TMC) from oxetane and CO2. Hence, based on these experimental ﬁndings, the formation of polycarbonate from the oxetane and CO2 coupling reaction was shown to occur via two different or concurrent pathways – that is, the intermediacy of TMC formation, and the subsequent polymerization and/or direct enchainment of oxetane and CO2 (Figure 8.16). The presence of small amounts of ether linkages in the copolymer also supported this conclusion. The proposed mechanism for the copolymerization of oxetane and CO2 is summarized in Scheme 8.9. Following the initial ring-opening and CO2 insertion into the resulting chromium–oxygen bond, two pathways are open for the intermediate:

• •

Pathway (1) involves the consecutive additions of oxetane and CO2 to generate the alternating copolymer. Pathway (2) leads to TMC formation by a back-biting process with ring closure.

Once TMC has been formed it can enter the polymer chain via a coordination– insertion process. In Scheme 8.9, the part of pathway (2) highlighted in red is highly dependent on the nature of the anionic leaving group. Indeed, for the good leaving groups (bromide and iodide ions) this pathway is competitive with oxetane enchainment, especially at lower temperatures, and provides a means of tuning the selectivity of the two pathways to either poly(TMC) or TMC formation.

8.4 Metal Catalysts for the Copolymerization of Oxetanes and CO2

0.4 0.35

Absorbance

0.3 0.25 0.2 0.15 0.1 0.05 0 0

200

400

600

800

1000

1200

1400

1600

Time (min)

Figure 8.15 Three-dimensional stack plot reaction proﬁle of the IR spectra collected every 3 min during the copolymerization

reaction of oxetane and CO2. The reaction was carried out at 383 K in toluene at 3.5 MPa CO2 pressure. Data taken from Ref. [64].

The advantage of proceeding exclusively via pathway (2) is the absence of ether linkages in the resultant copolymer, leading to improvements in its physical properties. The ﬁeld of metal-catalyzed copolymerization of oxetanes and CO2 will continue to ﬂourish, due not only to the versatility of the reaction but also to the aliphatic polycarbonate products being important components of thermoplastic elastomers that, in turn, have huge potential in medical applications such as sutures, drugdelivery systems, body, and dental implants, and tissue engineering. The exploration of other oxetane monomers (Figure 8.17) such as 3,3-dimethyloxetane and 3-methoxymethyl-3-methyloxetane, will surely provide a multitude of applications

233

8 Polymers from Carbon Dioxide

DHπ = 45.6 ± 3.01 kJ mol–1

DHπ = 74.08 ± 3.01 kJ mol–1

DSπ = -161.9 ± 8.21 J mol–1 K–1

DSπ = -72.28 ± 2.33 J mol–1 K–1

DGπ (383 K) = 107.6 kJ mol–1

Free Energy

234

O

+

DGπ (383 K) = 101.9 kJ mol–1

CO2

Reaction Coordinate Figure 8.16 Reaction coordinate diagram with activation parameters for the copolymerization of oxetane and CO2 and for the ring-opening polymerization of TMC.

N3

-

-

N3 CO2

Cr

Cr

O

O

O C O O

N3 N3 -

N3 Cr

+

CO2

(1)

O O

O Alternating Copolymer

N3

(2)

-

N3 O

Cr O

O

N3

O CO2

Scheme 8.9 Mechanistic aspects of the copolymerization of oxetane and CO2 catalyzed by (salen)CrN3/PPNN3.

8.5 Physical Methods for the Characterization of Copolymers Produced

O

O

3,3-dimethyloxetane Figure 8.17

OMe

3-methoxymethy-3-methyloxetane

Oxetane monomers.

for this methodology in the production of aliphatic polycarbonates that have diverse properties, and which can also be modiﬁed post polymerization.

8.5 Physical Methods for the Characterization of Copolymers Produced from Epoxides or Oxetane and Carbon Dioxide

Detailed analyses of the products and/or reaction solutions resulting from the copolymerization of epoxides or oxetanes and CO2 can be readily made by using 1 H NMR, 13C NMR, and IR spectroscopies. The molecular weights and polydispersities of the puriﬁed copolymers is generally achieved using gel-permeation chromatography (GPC) in tetrahydrofuran (THF) solution. Puriﬁcation of the copolymers is normally achieved by precipitation from a dichloromethane (DCM) solution with 1 M HCl in methanol, followed by vacuum drying. Alternative methods involving procedures for catalyst recycling are discussed later in the chapter. Typical NMR and IR spectroscopic data for the products and/or reaction solutions obtained from copolymerization processes of the most commonly employed epoxides (CHO or PO) or oxetane with CO2 are described in the following sections. 8.5.1 Cyclohexene Oxide Monomer

As seen in Figures 8.18–8.20, the presence of possible coupling products, namely polycarbonate and ﬁve-membered cyclic carbonate, can easily be assigned by using NMR and solution IR spectroscopies. With 1H NMR spectroscopy, the percent conversion to polymer can be monitored, based on the amount of cyclohexene monomer remaining in the reaction solution. Furthermore, the quantities of polycarbonate, ﬁve-membered cyclic carbonate, and ether linkages in the copolymer can be determined by integrating the peak area of the corresponding resonances at 4.62, 3.99, and 3.55 ppm, respectively. 13 C NMR spectroscopy of the reaction mixture is useful not only to determine the presence of coupling products, but also to assign the tacticity of the copolymer product. The solution IR spectrum of poly(cyclohexene carbonate) shows three stretching bands corresponding to νC=O of the copolymer at 1750 cm−1, and that of

235

8 Polymers from Carbon Dioxide

4.62 O O O

m

O

O

O

3.10

O

O

89.55

5.0

3.55

3.99

n

1.94

4.5

2.01

4.0

6.50

3.5

3.0

153.60

Figure 8.18 1H NMR spectrum in CDCl3 of the reaction solution obtained from the reaction between cyclohexene oxide and CO2.

Isotactic O

O

O

O

O

O

O

O

O

O

O O

O

O

154.04

O

Syndiotactic O

O

O

O

O O

O

O O

O

O O

153.03

O

O

O

154.83

236

156

155 13

154

153

152

Figure 8.19 C NMR spectrum in CDCl3 of the reaction solution obtained from the reaction between cyclohexene oxide and CO2.

1748.3

8.5 Physical Methods for the Characterization of Copolymers Produced

O O O

Absorbance

1.5

m O

O

O

0.5

0.0 1900

1850

1803.3

1819.7

1.0

1800

1750

1700

-1

Wavenumber(cm ) Figure 8.20 Infrared spectrum in CDCl3 of the reaction solution obtained from the reaction between cyclohexene oxide and CO2.

the ﬁve-membered cyclic carbonate at 1820 cm−1 and 1803 cm−1. The molecular weights of poly(cyclohexylene carbonate) obtained from CHO and CO2 in the presence of a (salen)CrN3/PPNN3 catalyst system, are generally ∼50 000, with PDI = 1.13. 8.5.2 Propylene Oxide Monomer

Propylene oxide represents a very attractive epoxide monomer for copolymerization with CO2, as poly(propylene carbonate) is industrially valuable. The low glass transition temperature (Tg) of 313 K, the sharp and clean decomposition above 473 K, and biodegradability of this copolymer are the reasons for its attracting interest in several applications. On a similar basis, 1H NMR spectroscopy is useful for assessing the coupling products resulting from the reaction of PO and CO2 (Figure 8.21). Likewise, 13C NMR represents a valuable means of monitoring the tacticity of the polycarbonate product; the 13C NMR of regio-irregular poly(propylene carbonate), with head-to-head (HH), tail-to-tail (TT), and head-to-tail (HT) junctions, is depicted in Figure 8.22. As seen in this ﬁgure, there is a much greater percentage of HT linkages compared to TT and HH, due to ring-opening occurring at the less-hindered carbon. The average molecular weights of poly(propylene carbonate) obtained from PO and CO2 in the presence of a (salen)CrN3/PPNN3 catalyst system was approximately 26 100, while the PDI was 1.11.

237

Hc

238

8 Polymers from Carbon Dioxide

CH3

O

CH3

Hc

O O

O n

Hb

O

O O

O

O

n

Hb

Ha

O

CH3

O

O

O

O CH3

Ha Hb

4.8

4.7

Hb

4.6

4.5

4.4

4.3

Hb

4.2

4.1

4.0

3.9

O n

CH3

Ha

Hc

3.8

Hc

H

3.7

3.6

3.5

3.4

H

3.3

1

Figure 8.21 H NMR spectrum in CDCl3 of the reaction solution obtained from the reaction between propylene oxide and CO2.

O

O

*

O

*

O

O O

O

HT

O O

TT

155.0

154.5

154.0

153.63

O

153.76

4.9

CH3

Ha

154.18 154.13 154.08

5.0

Hc

154.59

Ha

O

O

O O

*

O

O

HH

153.5

Figure 8.22 13C NMR spectrum in CDCl3 of regio-irregular poly(propylene carbonate) obtained from the reaction between propylene oxide and CO2.

O

8.5 Physical Methods for the Characterization of Copolymers Produced

8.5.3 Oxetane Monomer

Oxetane, a four-membered cyclic ether, is also an interesting monomer for the production of aliphatic polycarbonates, based on its coupling reaction with CO2, with both 1H NMR and IR spectroscopies being used to characterize the products obtained. In fact, 1H NMR spectroscopy has been used not only to assign and characterize the coupling products, but also to analyze the end groups of the polymers, which in part provide mechanistic information for the process [62]. The 1H NMR spectrum of a low-molecular-weight polycarbonate sample obtained from the copolymerization of oxetane and CO2, catalyzed by a (salen)CrCl/n-Bu4NN3 system, is shown in Figure 8.23. When the spectra were recorded in CDCl3, the copolymer exhibited two major signals at 4.23 ppm and 2.05 ppm, while ether linkages were observed in the copolymer at 3.50 ppm and 1.90 ppm. A –CH2OH end group was observed before and after puriﬁcation of the copolymer sample. These latter proton resonances appeared at 4.29 ppm, 3.73 ppm, and 1.90 ppm. The presence of an organic azide end group (–CH2N3) was also seen in the 1H NMR

a

a

O O

C

O

b

n

c O

n

d

b g

e

O C

O

f

OH

h O C O

N3 i O C O CH3

i e

g

4.5 1

4.0

c

d, f

h

3.5

3.0

2.5

2.0

Figure 8.23 H NMR spectrum in CDCl3 of puriﬁed poly(TMC) obtained from the reaction between oxetane and CO2.

239

240

8 Polymers from Carbon Dioxide

spectrum of the copolymer at 3.43 ppm, with the other two resonances being obscured by intense polymer signals at 4.43 and 2.05 ppm. The IR spectrum of this copolymer also exhibited an organic azide νN3 mode at 2100 cm−1 in tetrachloroethane. Finally, the polycarbonate displayed a 1H NMR resonance at 3.79 ppm, this being attributed to the –OC(O)OCH3 end group resulting from methanolysis of the original –OC(O)N3 end group following copolymer puriﬁcation from MeOH. The molecular weight values of poly(TMC) obtained from oxetane and CO2 in the presence of a (salen)CrCl/n-Bu4NN3 catalyst system were generally lower than the theoretical values (Mn = 11 050; Mn (theoretical) = 85 000). This situation was most likely due to a chain transfer mechanism arising from the presence of trace amounts of water in the system. However, when the catalytic runs were carried out under rigorously anhydrous conditions, the molecular weights more closely tracked the predicted values. When the spectra were recorded in CH2Cl2, the IR stretching bands of the carbonyl groups of poly(TMC) and TMC both appeared at 1750 cm−1, whereas in a mixture of toluene and oxetane they were separated by 20 cm−1; that is, the carbonyl group of TMC was seen at 1770 cm−1, and that of poly(TMC) at 1750 cm−1. By deconvoluting the IR spectra, it is possible to obtain the actual concentrations of both products, and hence monitor the reaction proﬁles for both poly(TMC) and TMC formation and/or consumption (Figure 8.24). Finally, another method commonly used to characterize these copolymeric materials is that of matrix-assisted laser desorption ionization-time of ﬂight (MALDI-TOF) mass spectrometry. When using this technique, it is possible to investigate the molecular weight distributions, end groups, and polymer constitution of low-molecular-weight samples [63]. MALDI-TOF may also provide vital

Figure 8.24 Selected IR spectrum of a reaction solution obtained from a reaction between oxetane and CO2. (a) Undeconvoluted IR spectrum; (b) Deconvoluted IR spectrum corresponding to poly(TMC); (c) Deconvoluted IR spectrum corresponding to TMC.

8.6 Copolymer Isolation and Catalyst Recycling

information regarding the initiation, propagation, and termination steps of the polymerization reaction.

8.6 Copolymer Isolation and Catalyst Recycling

A ﬁnal issue that should be addressed in the metal-catalyzed production of copolymers from CO2 and epoxides or oxetanes is that of separating the catalyst from the reaction mixture. As noted above, the traditional method of isolating a puriﬁed copolymer is by the addition of acidiﬁed methanol (1 M HCl in MeOH) to a DCM solution of the polymer/catalyst mixture. This results in precipitation of the “puriﬁed” polymer, while the catalyst and any other byproducts remain in solution. Although the polymer can then be ﬁltered and dried, some catalyst material will inevitably still be retained within in the polymeric material. Whilst this process can be repeated several times in order to achieve a more complete removal of any metal residues, it will normally result in decomposition of the catalyst such that it cannot be recycled. Because the clean-up costs of this process are potentially very high, many investigations have been conducted into ﬁnding an alternative to using large volumes of extraneous environmentally harmful solvents, but rather to use recyclable catalyst systems. One approach to solving this dilemma would be to modify the catalyst structure such that it would aid in any separation, but not inhibit the activity of the catalyst towards polymer production. For example, Bergbreiter and coworkers synthesized a chromium salen complex where the ligand was substituted with ∼1000 Da poly(isobutylene) (PIB) on the salen ligand (Figure 8.25), instead of the traditional tert-butyl groups [64]. This catalyst showed almost identical reactivity to comparable Cr(salen) complexes, and had the added bonus of easy separation. Upon completion of the reaction, the polymer/catalyst mixture was dissolved in MeCN to which a small portion of concentrated HCl was added with stirring. After concentration, heptane was added, and the polymer and catalyst were separated and isolated. The Cr(PIBsalen) catalyst provided not only a more “green” route to polymer/catalyst separation, but also proved to be about fourfold more efﬁcient at removing the Cr catalyst from the polymer than the acid/methanol method (Figure 8.26). When the catalyst was tested for recyclability, it was found to have lost

Figure 8.25

Cr(PIBsalen)Cl utilized for the copolymerization of CO2 and cyclohexene oxide.

241

242

8 Polymers from Carbon Dioxide

(a)

(b)

Figure 8.26 Colorimetric comparison of polycarbonate obtained: (a) via acidiﬁed methanol puriﬁcation; (b) utilizing the Cr(PIBsalen)Cl/CH3CN/heptane system.

Figure 8.27

Co(salen) complex utilized by Lee for the copolymerization of CO2 and PO [54].

∼20–30% of its activity after one cycle, most likely because its exposure to concentrated acid during the recycling scheme had caused it to degrade. Another example of catalyst modiﬁcation was achieved by synthesizing a Co(salen)-based catalyst with pendent ammonium salts (Figure 8.27) [54]. This complex was able to catalyze the copolymerization of CO2 and PO at TOF-values of up to 26 000 h−1, with high molecular weights and low polydispersities. Upon completion of the reaction the polymer/catalyst mixture was ﬁltered through a pad of silica which yielded the puriﬁed polymer product. When the Co(salen) complex had been recovered from the silica it was recycled several times, with little to no loss of activity. A variety of homogeneous and heterogeneous catalyst systems have been designed for the more efﬁcient separation of metal–salen complexes from product/ catalyst mixtures [65]. One approach has been the addition of highly ﬂuorinated groups to the salen ligand so as to facilitate solubility in both liquid and scCO2. Another approach has been to add imidazolium salts to the salen ligand’s architecture so as to facilitate separation via ionic liquids. Along with the previous examples, salen–metal complexes have been tethered to several different solid supports (e.g., silica, alumina, polystyrene resin beads) to produce heterogeneous catalysts that can simply be isolated by ﬁltration at the end of the reaction, and then recycled.

8.6 Copolymer Isolation and Catalyst Recycling

Scheme 8.10 Formation of switchable polarity carbamate solvent.

epoxide & CO2 PPNX

P O Cr

NHEtBu

X

O

P O Cr

O Cr

X

X

NEtBu

distillation

CO2 O

NHEtBu

O

O NEtBu

O

Cr

Cr

X

X

filtration

NEtBu P HO

P HO Scheme 8.11 Recycling of catalyst and SPS following the copolymerization of CO2 and CHO. Colors indicate species in low-polarity SPS (green), high-polarity SPS (yellow), and without solvent (colorless).

Another approach to polymer/catalyst separation is to replace the use of costly wasteful solvents with a procedure that produces little to no waste, and requires no catalyst modiﬁcation. These switchable polarity solvents (SPS), as developed by Jessop and coworkers, are able to utilize the reaction of secondary amines with CO2 to reversibly form carbamate salt ionic liquids (Scheme 8.10) [66]. Upon the completion of the polymerization reaction, EtBuNH (Et = ethyl, Bu = n-butyl) is added to dissolve the polymer/catalyst mixture. The subsequent bubbling of CO2 through this solution results in a precipitation of the polymer product, such that the puriﬁed polymer can be isolated by simple ﬁltration. The catalyst can then be recovered by distilling the amine/carbamate mixture, and the entire system can (in theory) be recycled (Scheme 8.11). More comprehensive discussions of considerations relevant to the development of viable industrial processes for the catalytic conversion of epoxides and CO2 are included in the reviews of Keurentjes and coworkers [67], and Luinstra [68].

243

244

8 Polymers from Carbon Dioxide

8.7 Copolymerization of Carbon Disulﬁde and Epoxides and Episulﬁdes

The addition of sulfur atoms can lead to copolymers with desirable properties over their oxygen-only analogues, namely polycarbonates. Sulfur-rich polymers are wellknown for their thermal stability, optical properties, and potential use as heavymetal scavengers. Much of the research surrounding the formation of poly (trithiocarbonates) has centered on the ROP of cyclic thiocarbonates and the utilization of di-thiols and/or thiophosgene [19c, 69]. Some investigations have been conducted, however, into identifying alternative routes to poly(trithiocarbonates) that are analogous to the copolymerization of CO2 and epoxides. Metal complexes, such as zinc–cobalt double metal cyanides (DMCs) and chromium–salen complexes, have been employed as catalysts for the formation of copolymers [70]. The production of an all-sulfur polycarbonate analogue was achieved via the metal-catalyzed copolymerization of carbon disulﬁde and a thiorane [71]. Additionally, the copolymerization of epoxides and carbon disulﬁde has been accomplished (Equation 8.6) [70]. O

S +

C

A

B

cat. C

S

(8.6) n

A,B,C = S or O One complication in this area is the S/O rearrangement that occurs during the metal-catalyzed copolymerization reaction. It would be expected that a reaction between carbon disulﬁde and an epoxide would result in the corresponding poly(dithiocarbonate) (A = O, B,C = S); however, the observation of several different combinations of sulfur and oxygen have been observed. This “atomscrambling” process is seen at all reaction temperatures, and can also be affected by reaction pressure. The addition of sulfur to aliphatic copolymers results in polymeric materials with higher Tg than their all-oxygen counterparts. In the case of the sulfur analogue of bisphenol A–polycarbonate (BPA-PC), the addition of sulfur resulted in a lowering of the Tg, but caused increases in both the crystallinity and melting point of the polymer [69c]. Another method of creating these polymeric materials is via the ROP of cyclic thiocarbonates. Such starting materials can be formed from the reaction between carbon disulﬁde and epoxides, and also as byproducts from the copolymerization process. It is worth mentioning here that the same S/O rearrangement is observed during the synthesis of the cyclic materials as in the production of poly(thiocarbonate) products. Unfortunately, the ROP requires a high temperature and is plagued by the formation of polyether and polythioether linkages in the polymeric material. Hence, it has proved to be an ineffective route for the production of highmolecular-weight, ether-free polymers.

8.9 Concluding Remarks

8.8 Copolymers from Aziridines and Carbon Dioxide

Aziridines represent another group of cyclic monomers that are capable of copolymerizing with CO2 to potentially provide useful polymeric materials, namely polyurethanes. Early studies of the reactions of aziridines with CO2 led to the production of cyclic urethanes [72] and polymers [73, 74], but the polymeric product was shown to consist of both urethane and amine linkages, as depicted in Equation 8.7. However, because the rate of homopolymerization of aziridines is similar to that of the copolymerization of aziridines and CO2, the urethane linkages were limited to ∼30%. O

H N

H N

C + CO2

N H

(8.7)

O n

m

Although, to date, there has been a lack of progress in this area in general, the most recent signiﬁcant contribution has derived from the laboratory of Kayaki and Ikariya [75]. These authors examined the reaction depicted in Equation 8.7 under scCO2 conditions, using NMR spectroscopy, and showed that an increase in CO2 pressure and a concomitant increase in CO2 density greatly enhanced the urethane content of the copolymer. For example, at a temperature of 373 K, upon varying the CO2 pressure from 3.0 to 22.0 MPa, the urethane content was increased from 32% to 62%, with corresponding molecular weights (Mw) of 6.5 × 103 and 2.7 × 104, respectively. Furthermore, it was shown that the addition of 3.6 wt% N,N-dimethylacetamide and a pressure of 22.0 MPa led to an improvement of the urethane content to 0.74.

8.9 Concluding Remarks

In this chapter, some of the essential aspects of the synthesis and characterization of copolymers derived from the coupling of CO2 with various monomers, namely, epoxides, oxetanes, and aziridines, have been reviewed. In addition, the use of carbon disulﬁde as a resource for copolymer production was introduced, and the present understanding of the mechanistic aspects of processes involving cyclic ethers and CO2 catalyzed by well-deﬁned metal systems has been emphasized. This knowledge has led to the development of catalytic systems capable of controlling not only the product selectivity but also the regio- and stereoregularities of the resultant copolymers. The transformation of CO2 into polymeric materials, thus offsetting the demand for petroleum-based resources, has matured to a point where it is currently being

245

246

8 Polymers from Carbon Dioxide

used in speciﬁc instances, albeit on a limited industrial scale. Today, CO2-based polymers such as poly(ethylene carbonate) and poly(propylene carbonate) are produced commercially in the United States by Empower Materials and Novomer, Inc. It is anticipated that interest in this area of research will intensify in the future, with attention becoming focused on the production of other copolymers from CO2 and a more diversiﬁed group of comonomers. Fundamental studies of the type described in this chapter hopefully will provide the knowledge and stimulus to develop the large-scale production of valuable polymeric materials via these “greener” processes.

Acknowledgments

The authors’ original research on the utilization of CO2 as a source of chemical carbon has been funded over the years by the US. National Science Foundation and the R.A. Welch Foundation of Texas.

References 1 Brunelle, D.J. and Korn, M.R. (eds) (2005) Advances in Polycarbonates. Proceedings of Symposium of the American Chemical Society held March 2003 in Washington, DC, ACS Symp. Ser. 898, p. 281. 2 Clements, J.H. (2003) Ind. Eng. Chem. Res., 42, 663. 3 Pell, A.S. and Pilcher, G. (1965) Trans. Faraday Soc., 61, 71. 4 (a) Seal, B.L., Otero, T.C., and Panitch A. (2001) Mater. Sci. Eng., R34, 147. (b) Beiser, I.H. and Konat, I.O. (1990) J. Am. Podiatr. Med. Assoc., 80, 272. 5 Marianucci, E., Berti, C., Pilati, F., Manaresi, P., Guaita, M., and Chiantore, O. (1994) Polymer, 35, 1564. 6 Stevens, H.C. (1966) US Patent 3,248,415, Pittsburgh Plate Glass Company. 7 Inoue, S., Koinuma, H., and Tsuruta, T. (1969) J. Polym. Sci., Part B: Polym. Phys., 7, 287. 8 Inoue, S., Koinuma, H., and Tsuruta, T. (1969) Makromol. Chem., 130, 210. 9 Rokicki, A. and Kuran, W. (1981) J. Macromol. Sci., Rev. Macromol. Chem., C21, 135. 10 For a more comprehensive coverage of this topic see the recent reviews by: (a) Coates, G.Q. and Moore, D.R. (2004) Angew. Chem., Int. Ed., 43, 6618.

11 12

13 14

15 16

17

18 19

(b) Sugimoto, H. and Inoue, S. (2004) J. Polym. Sci., Part A: Polym. Chem., 42, 5561. (c) Inoue, S., Koinuma, H., and Tsuruta, T. (1971) US Patent 3,585,168. Soga, K., Imai, E., and Hattori, I. (1981) Polym. J., 13, 407. Darensbourg, D.J., Holtcamp, M.W., and Reibenspies, J.H. (1996) Polyhedron, 15, 2341. Zheng, Y.-Q., Liu, J.-L., and Zhang, H.-L. (2000) Z. Kristallogr., 215, 535. (a) Rokicki, A. (1990) US Patent 4,943,677 (Air Products and Chemicals, Inc., and Arco Chemicals Co.). (b) Motika, S. (1991) US Patent 5,026,676 (Air Products and Chemicals, Inc., Arco Chemicals Co., and Mitsui Petrochemical Industries Ltd.). www.empowermaterials.com (accessed 20 August 2008). Darensbourg, D.J., Stafford, N.W., and Katsurao, T. (1995) J. Mol. Catal. A: Chemical, 104, L1. Coates, G.W. and Moore, D.R. (2004) Angew. Chem., Int. Ed., 43, 6618. Darensbourg, D.J. (2007) Chem. Rev., 107, 2388. (a) Chisholm, M.H. and Zhou, Z. (2004) J. Mater Chem., 14, 3081.

References

20 21 22 23

24

25

26 27

28

29

30 31

32

33

34

(b) Darensbourg, D.J. and Holtcamp, M.W. (1996) Coord. Chem. Rev., 153, 155. (c) Sugimoto, H. and Inoue, S. (2004) J. Polym. Sci., Part A: Polym. Chem., 42, 5561. (d) Darensbourg, D.J., Mackiewicz, R.M., Phelps, A.L., and Billodeaux, D.R. (2004) Acc. Chem. Res., 37, 836. (e) Ochiai, B. and Endo, T. (2005) Prog. Polym. Sci., 30, 183. Aida, T., Ishikawa, M., and Inoue, S. (1986) Macromolecules, 19, 8. Chisholm, M.H. and Zhou, Z. (2004) J. Am. Chem. Soc., 126, 11030. Kruper, W.J. and Dellar, D. (1995) J. Org. Chem., 60, 725. Mang, S., Cooper, A.I., Colclough, M.E., Chauhan, N., and Holmes, A.B. (2000) Macromolecules, 33, 303. (a) Darensbourg, D.J. and Holtcamp, M.W. (1995) Macromolecules, 28, 7577. (b) Darensbourg, D.J., Holtcamp, M.W., Struck, G.E., Zimmer, M.S., Niezgoda, S.A., Rainey, P., Robertson, J.B., Draper, J.D., and Reibenspies, J.H. (1999) J. Am. Chem. Soc., 121, 107. Super, M., Berluche, E., Costello, C., and Beckman, E.J. (1997) Macromolecules, 30, 368. Cheng, M., Lobkovsky, E.B., Coates, G.W. (1998) J. Am. Chem. Soc., 120, 11018. Allen, S.D., Moore, D.R., Lobkovsky, E.B., and Coates, G.W. (2002) J. Am. Chem. Soc., 124, 14284. Byrne, C.M., Allen, S.D., Lobkovsky, E.B., and Coates, G.W. (2004) J. Am. Chem. Soc., 126, 11404. Moore, D.R., Cheng, M., Lobkovsky, E.B., and Coates, G.W. (2003) J. Am. Chem. Soc., 125, 11911. Xiao, Y., Wang, Z., and Ding, K. (2006) Macromolecules, 39, 128. Darensbourg, D.J., Rodgers, J.L., Mackiewicz, R.M., and Phelps, A.L. (2004) Catal. Today, 98, 485. Darensbourg, D.J., Mackiewicz, R.M., Rodgers, J.L., Fang, C.C., Billodeaux, D.R., and Reibenspies, J.H. (2004) Inorg. Chem., 43, 6024. Li, B., Wu, G.-P., Ren, W.-M., Wang, Y.-M., Rao, D.-Y., and Lu, X.-B. (2008) J. Polym. Sci., Part A: Polym. Chem., 46, 6102. (a) Darensbourg, D.J. and Fitch, S.B. (2007) Inorg. Chem., 46, 5474.

35

36 37

38 39 40 41

42 43

44

45 46 47

48

49 50

51

52

(b) Darensbourg, D.J. and Fitch, S.B. (2008) Inorg. Chem., 47, 11868. Darensbourg, D.J., Mackiewicz, R.M., and Billodeaux, D.R. (2005) Organometallics, 24, 144. Cohen, C.T., Chu, T., and Coates, G.W. (2005) J. Am. Chem. Soc., 127, 10869. (a) Aida, T., Ishikawa, M., and Inoue, S. (1986) Macromolecules, 19, 8. (b) Sugimoto, H., Ohtsuka, H., and Inoue, S. (2005) J. Polym. Sci., Part A: Polym. Chem., 43, 4172. (c) Nakano, K., Kamada, T., and Nozaki, K. (2006) Angew. Chem. Int. Ed., 45, 7274. (d) Qin, Y., Wang, X., Zhang, S., Zhao, X., and Wang, F. (2008) J. Polym. Sci., Part A: Polym. Chem., 46, 5959. Darensbourg, D.J. and Fitch, S.B. (2009) Inorg. Chem., 48, 8668. Darensbourg, D.J. and Moncada, A.I. (2008) Inorg. Chem., 47, 10000. Weil, M. and Khalaji, A.D. (2008) Anal. Sci., 24, x19. Darensbourg, D.J. and Yarbrough, J.C. (2002) J. Am. Chem. Soc., 124, 6335. Darensbourg, D.J. and Phelps, A.L. (2005) Inorg. Chem., 44, 4622. Eberhardt, R., Allmendinger, M., and Rieger, B. (2003) Macromol. Rapid Commun., 24, 194. Qin, Z., Thomas, C.M., Lee, S., and Coates, G.W. (2003) Angew. Chem., Int. Ed., 42, 5484. Lu, X.-B. and Wang, Y. (2004) Angew. Chem., Int. Ed., 43, 3574. Paddock, R.L. and Nguyen, S.T. (2005) Macromolecules, 38, 6251. Lu, X.-B., Shi, L., Wang, Y.-M., Zhang, R., Zhang, Y.-J., Peng, X.-J., Zhang, Z.-C., and Li, B. (2006) J. Am. Chem. Soc., 128, 1664. Cohen, C.T. and Coates, G.W. (2006) J. Polym. Sci., Part A: Polym. Chem., 44, 5182. Sugimoto, H. and Kuroda, K. (2008) Macromolecules, 41, 312. Darensbourg, D.J., Bottarelli, P., and Andreatta, J.R. (2007) Macromolecules, 40, 7727. Luinstra, G.A., Haas, G.A., Molnar, F., Bernhart, V., Eberhardt, R., and Rieger, B. (2005) Chem. Eur. J., 11, 6298. Nakano, K., Kamada, T., and Nozaki, K. (2006) Angew. Chem., Int. Ed., 45, 7274.

247

248

8 Polymers from Carbon Dioxide 53 Noh, E.K., Na, S.J., Sujith, S., Kim, S.-W., and Lee, B.Y. (2007) J. Am. Chem. Soc., 129, 8082. 54 Sujith, S., Min, J.K., Seong, J.E., Na, S.J., and Lee, B.Y. (2008) Angew. Chem., Int. Ed., 47, 7306. 55 Darensbourg, D.J., Ulusoy, M., Karroonnirun, O., Poland, R.R., Reibenspies, J.H., and Çetinkaya, B. (2009) Macromolecules, 42, 6992. 56 Koinuma, H. and Hirai, H. (1977) Makromol. Chem., 178, 1283. 57 Koinuma, H. and Hirai, H. (1977) Makromol. Chem., 178, 241. 58 Baba, A., Meishou, H., and Matsuda, H. (1984) Makromol. Chem. Rapid Commun., 5, 665. 59 Baba, A., Kashiwagi, H., and Matsuda, H. (1985) Tetrahedron Lett., 26, 1323. 60 Baba, A., Kashiwagi, H., and Matsuda, H. (1987) Organometallics, 6, 137. 61 Darensbourg, D.J., Ganguly, P., and Choi, W. (2006) Inorg. Chem., 45, 3831. 62 Darensbourg, D., Moncada, A.I., Choi, W., and Reiebenspies, J.H. (2008) J. Am. Chem. Soc., 130, 6523. 63 Duchateau, R., Van Meerendonk, W.J., Yajjou, L., Staal, B.B.P., Koning, C.E., and Gruter, G.-J. (2006) Macromolecules, 39, 7900. 64 Hongfa, C., Tian, J., Andreatta, J., Darensbourg, D.J., and Bergbreiter, D.E. (2008) Chem. Commun., 975.

65 Baleizão, C. and Garcia, H. (2006) Chem. Rev., 106, 3987. 66 Pahn, L., Andreatta, J.R., Horvey, L.K., Edie, C.F., Luco, A.-L., Mirchandani, A., Darensbourg, D.J., and Jessop, P.G. (2008) J. Org. Chem., 73, 127. 67 van Schilt, M., Kemmere, M., and Keurentjes, J. (2006) Catal. Today, 115, 162. 68 Luinstra, G.A. (2008) Polym. Rev., 48, 192. 69 (a) Choi, W., Sanda, F., and Endo, T. (1998) Macromolecules, 31, 2454. (b) Kricheldorf, H.R. and Damrau, D.-O. (1998) Macromol. Chem. Phys., 199, 2589. (c) Berti, C., Celli, A., and Marianucci, E. (2002) Eur. Polym. J., 38, 1281. (d) Berti, C., Marianucci, E., and Pilati, F. (1988) Makromol. Chem., 189, 1323. 70 Zhang, X.H., Liu, F., Sun, X.K., Chen, S., Du, B.Y., Qi, G.R., and Wan, K.M. (2008) Macromolecules, 41, 1587. 71 Nakano, K., Tatsumi, G., and Nozaki, K. (2007) J. Am. Chem. Soc., 129, 15116. 72 Soga, K., Hosoda, S., Nakamura, H., and Ikeda, S. (1976) J. Chem. Soc. Chem. Commun., 16, 617. 73 Soga, K., Chiang, W.Y., and Ikeda, S. (1974) J. Polym. Sci. Polym. Chem. Ed., 12, 121. 74 Inoue, S. (1976) Chemtech, 588. 75 Ihata, O., Kayaki, Y., and Ikariya, T. (2004) Angew. Chem., Int. Ed., 43, 717.

249

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane for Synthesis Gas Production, Using the Tapered Element Oscillation Microbalance Wie Pan and Chunshan Song

9.1 Introduction

Carbon dioxide (CO2) can be reduced and utilized as carbon monoxide (CO), methane (CH4), and methanol (CH3OH), as well as other chemicals and materials [1a–d]. The reaction for the CO2 reforming of methane converts CO2, via methane, into CO and H2. This reaction, which initially was proposed as a means of using CO2 to produce industrially useful CO and H2 [1e], is also viewed as a way of reducing the emissions of CO2, as a greenhouse gas. Unfortunately, the application of CO2 reforming of methane has been much less successful than the steam reforming process, because catalysts in CO2 reforming have a limited lifetime due to problems of severe carbon deposition that lead to catalyst disintegration and deactivation. In addition, CO2 reforming is a more endothermic reaction compared to steam reforming, and requires a signiﬁcant energy input. In fact, this reaction would require a supply of CO2 which is not readily available, as well as the separation and recovery of CO2 on-site in manufacturing plants. On the other hand, CO2 reforming can be used for converting and utilizing CO2-rich natural gas [1f], notably as some natural gas resources contain up to 50 vol% CO2; however, these gases have not yet been utilized commercially due to their high CO2 concentrations. Furthermore, CO2 reforming can be used to convert the ﬂue gas of coalﬁred or natural gas-ﬁred electric power plants by the tri-reforming process, as recently proposed and established at the Pennsylvania State University [1f, g]. Tri-reforming is a synergetic combination of CO2 reforming, steam reforming, and the partial oxidation of methane [1f]. CO2 reforming may also be used for the conversion and utilization of bio-gas, such as anaerobic digester gas and landﬁll gas which contain up to 40–50% CO2 in addition to methane. The present studies were initiated to investigate the CO2 reforming of methane, with attention focused on the problem of carbon formation during the reaction process. The system of CO2 reforming of methane incorporates the same ﬁve gases (CO2, CH4, H2O, CO, H2) as in the system of steam reforming; the related reactions are shown in Equations 9.1 to 9.8. The major difference between these two systems Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

250

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

is the relative concentration of each component in the two reaction processes. If it is assumed that both steam reforming and CO2 reforming begin with reactant gas mixtures of H2O/CH4 = 1 and CO2/CH4 = 1, that the CH4 conversions are all 80%, and that the water gas shift reaction and other side reactions are not considered, then the gas mixture after steam reforming will contain about 22% CO, 67% H2, 5.5% CH4, and 5.5% H2O, while the gas mixture after CO2 reforming will contain 44% CO, 44% H2, 5.5% CO2, and 5.5% CH4. One of the major problems in CO2 reforming is the formation of carbon deposits [2, 3]. In the CO2 reforming reaction, the possible routes of carbon formation include methane decomposition (Equation 9.6) and/or CO disproportionation (Equation 9.7). The clariﬁcation of this issue is necessary because it will help to understand the mechanism of carbon formation in this reforming reaction. Such insight will also be beneﬁcial not only for the design of reactors but also for the creation of stable and carbon-resistant catalysts.

•

CO2 reforming of methane: CO2 + CH4 = 2 CO+2 H2

•

ΔH °= 247.3 kJ mol −1

Steam reforming of methane: H2O + CH4 = CO + 3 H2

•

ΔH °= 206.3 kJ mol −1

2

4

2

ΔH °= −35.6 kJ mol −1

(9.5)

ΔH °= −172 kJ mol −1

(9.6)

Methane decomposition: CH4 = C + 2 H2

•

ΔH ° = −41 kJ mol −1

CO disproportionation: 2 CO = C + CO2

•

(9.4)

Water gas shift reaction: CO + H2O = CO2 + H2

•

(9.3)

Complete oxidation of methane: 2 O2 + CH4 = CO2 + 2 H2O ΔH ° = −880 kJ mol −1

•

(9.2)

Partial oxidation of methane:

( 21 )O + CH = CO + 2 H •

(9.1)

ΔH ° = 75 kJ mol −1

(9.7)

Deoxygenation of CO: CO + H2 = C + H2O ΔH °= −131 kJ mol −1

(9.8)

When Wang and Lu [2] studied the dependence of the carbon formation rate (Rc) on the partial pressure of CH4 (PCH4) and CO2 (PCO2) in CO2 reforming of methane, the reaction order of carbon formation rate with respect to CO2 partial pressure (shown in Equation 9.9) was seen to be negative, which indicated that CO2 in the feed would inhibit carbon deposition. Hence, Wang and Lu [2] concluded that

9.1 Introduction

methane decomposition was the main route of carbon deposition in CO2 reforming over Ni/γ-Al2O3 catalysts. Rc = K cPCH 41.18PCO2 −0.68

(9.9)

By analyzing their measurement conditions, it is estimated that the carbon formation rates were measured under differential reaction conditions [space velocity ca. 1 440 000 h−1 and contact time (W/F) ca. 0.015 g h mol−1], although the methane conversion was not mentioned in the report. As there was very little CO present in the system studied by Wang and Lu, the effect of CO on carbon formation could not have been fully taken into consideration. In addition, if methane decomposition were to be the main route of carbon formation in CO2 reforming, it would be difﬁcult to explain the experimental phenomenon observed by Fujimoto and coworkers [3], that carbon formation increased in line with increases in reaction pressures (from 0.1 to 2.0 MPa). Thermodynamically, high reaction pressures do not favor methane decomposition in itself. Alternatively, others [4–7] have suggested that CO disproportionation is responsible for carbon formation in CO2 reforming. As CO disproportionation is an exothermic reaction, the equilibrium carbon formation will increase with increasing pressures, and decrease with increasing temperatures. This explains clearly the pressure effect on carbon formation observed by Fujimoto and coworkers [8], and also agrees with the “temperature effect” observed by Zhang and Verykios [6] and Richardson and Paripatyadar [5]. These groups observed that the amount of carbon on Ni/γ-Al2O3 and Ru or Rh/Al2O3 after a 2 h or 8 h reaction decreased with the increase of reaction temperatures, even starting from 773 K or 883 K. Fujimoto and coworkers [8] reported that carbon formation after a 4 h reaction over NiO– MgO solid solution catalysts decreased when the temperature was raised from 1073 K to 1173 K. Nonetheless, there is barely any direct experimental evidence to prove that CO is responsible for carbon formation in CO2 reforming. Swann et al. [9] used isotopic labeling and temperature-programmed oxidation (TPO) (ex situ) approaches to show that the carbon which had formed during the CO2 reforming of methane may have derived from both CH4 and CO2. On the basis of investigations conducted in their laboratory, the present authors believe that the methods and conditions used to study carbon formation are critical in order to correctly understand the behavior of carbon formation. For example, the method used to measure carbon formation, whether in situ or ex situ, may affect the results, with ex situ techniques not providing complete kinetic information. As another example, in a ﬁxed-bed ﬂow reactor, when the CO2 reforming reaction approaches equilibrium the gas-phase compositions along the catalyst bed are different; close to the outlet of the catalyst bed the gas phase contains not only unconverted CH4 and CO2, but also products such as CO, H2. The appearance of CO and H2 together with CH4 and CO2 in the gas phase, might affect the carbon formation behavior and carbon morphology [10, 11]. It is possible that some experiments [2] did not reﬂect the carbon formation behavior in real CO2 reforming processes because they were conducted at very low CH4 conversions, with almost

251

252

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

no CO present in the gas phase. Clearly, in order to study carbon formation in CO2 reforming, the gas-phase composition should be close to that in a reformer containing both CH4 and CO. For the studies described in this chapter, a novel system was employed, namely the tapered element oscillation microbalance (TEOM), to monitor the dynamic process of carbon formation under real CO2 reforming conditions. The use of TEOM enables carbon formation to be measured in situ, and kinetic studies of carbon formation to be performed at different partial pressures of CH4 and CO.

9.2 Thermodynamic Analysis of Carbon Formation from CH4 or CO

As both CH4 and CO represent possible sources of carbon formation during the CO2 reforming reaction, the thermodynamic behavior of carbon formation from CH4 decomposition and CO disproportionation was ﬁrst studied. CH4 decomposition is an endothermic reaction, with carbon formation being favored at higher temperatures (see Figure 9.1). In contrast, CO disproportionation is an exothermic reaction, such that both equilibrium CO conversion and carbon formation decline as the reaction temperature is increased (Figure 9.2). However, in practice CH4 decomposition and CO disproportionation are often conducted over catalysts due to the kinetics of these reactions. It has been found that carbon formed in CH4 decomposition and CO disproportionation over catalysts may not be in the form of graphite, and indeed may have a higher free energy compared to graphite [12, 13]. The difference in the free energy between carbon on catalysts and graphite becomes smaller when carbon on catalysts is obtained at higher reaction temperatures (>873 K). In order to demonstrate the effect of the thermodynamic properties of carbon on catalysts on the equilibrium of methane decomposition and CO disproportionation, the free energy data of carbon on Ni catalysts, as reported by Dent [12] is used for the calculations. Subsequently, the results obtained were compared with those calculated based on graphite (Figures 9.1 and 9.2). The comparison revealed that the high free energy of carbon on Ni catalysts resulted in a lower conversion of CH4 and CO, as well as a reduced carbon formation in both the CH4 decomposition and CO disproportionation reactions. For example, the biggest difference in CH4 conversion and carbon formation was observed at 773 K in the CH4 decomposition reaction. At this temperature, the equilibrium CH4 conversion and carbon formation from 1 mol of methane based on graphite were approximately 28.6% and 0.286 mol, respectively, whereas the equilibrium CH4 conversion and carbon formation based on carbon on Ni catalysts was only 16.7% and 0.167 mol, respectively. However, with an increase in temperature these differences gradually disappeared as the free energy of carbon on Ni catalysts became close to that of graphite.

9.2 Thermodynamic Analysis of Carbon Formation from CH4 or CO (a)

CH4 conversion (%)

100

Graphite

80

Carbon on Ni

60

40

20

0

273

Carbon formation (mol/mol methane)

(b)

473

673

873

1073

1273

1073

1273

Temperature (K) 1

0.8

Graphite Carbon on Ni

0.6

0.4

0.2

0

273

473

673

873

Temperature (K) Figure 9.1 Difference of CH4 conversion and carbon formation in CH4 = C (graphite) + 2H2 and CH4 = C (carbon on nickel) + 2H2. The free energy of carbon on a nickel catalyst was derived from Ref. [12].

Based on these observations, the decision was taken to use the thermodynamic properties of graphite in the thermodynamic analysis in CO2 reforming, because this reaction has applicable conversions only at temperatures above 973 K. At this point, the difference in free energy between graphite and carbon on catalysts becomes so small that it has a negligible effect on the thermodynamic analysis.

253

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane (a) 100

CO conversion (%)

Graphite Carbon on Ni

80 60 40 20 0

273

473

673

873

1073

1273

Temperature (K)

(b) 1

Carbon formation (mol/mol CO)

254

Graphite

0.8

Carbon on Ni

0.6

0.4

0.2

0

273

473

673

873

1073

1273

Temperature (K) Figure 9.2 Difference of CO conversion and carbon formation in 2CO = C (graphite) + CO2 and 2CO = C (carbon on nickel) + CO2. The free energy of carbon on a nickel catalyst was derived from Ref. [12].

9.3 Thermodynamic Analysis of Carbon Formation in CO2 Reforming of Methane

Figure 9.3 shows the equilibrium carbon formation in CO2 reforming of methane. Instead of analyzing each individual carbon-formation reaction, the thermodynamic analysis considers all possible reactions that might occur during the real CO2 reforming system. In a practical CO2 reforming system, the CO2–CH4 reforming reaction and side reactions, such as water gas shift reaction and carbon formation reactions, take place simultaneously. Hence, in order to obtain an equilibrium composition, the Gibbs energy minimization method was employed

9.3 Thermodynamic Analysis of Carbon Formation in CO2 Reforming of Methane

Carbon formation (mol/mol methane)

2.00

1.50

1.00

0.50

0.00

273

473

673

873

1073

1273

Temperature (K) Figure 9.3

Theoretical calculation of carbon formation in CO2 reforming of methane.

[14] and minimization was carried out using a commercial HSC Chemistry program. In a CO2 reforming reaction system, there exist at least six components, namely CO2, CH4, CO, H2, H2O, and C. Based on the mass balance of these three elements (C, H, and O), three equations can be established. In addition, according to the rule described by Denbigh [14], the number of independent reactions in a reaction system can be determined by writing down the formation equations of all compounds in the reaction system from their elemental atoms, followed by combining these equations in such a way as to eliminate any free atoms which are not actually present. As a result, in the CO2 reforming system, there are three independent reactions which can be expressed as three equilibrium equations. Therefore, in theory, the equilibrium composition in the CO2 reforming system can also be obtained by solving the six equations. It is seen from Figure 9.3 that the trend of carbon formation with the increase of temperature in CO2 reforming is similar to that observed in Figure 9.2. A higher carbon formation at low temperatures, and a lower carbon formation at higher temperatures, suggest that CO disproportionation probably dictates equilibrium carbon formation in the CO2 reforming reaction. This seems reasonable because CO is the dominant source to produce carbon at lower temperatures. The CO disproportionation reaction has large equilibrium constant towards carbon formation at low temperatures, due to the highly exothermic property of this reaction. At higher temperatures, most of the CH4 may react with CO2 and form CO and, as a result, very little methane is left. Hence, CO will still largely determine the equilibrium carbon formation at higher temperatures in a CO2 reforming reaction.

255

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

9.4 TEOM Measurement of Carbon Formation in CO2 Reforming of Methane

Carbon formation in the CO2 reforming of methane with CH4/CO2 (molar ratio) = 1 was measured using the TEOM at 923 K and 0.1 MPa over 25 mg ICI catalyst. The detailed TEOM experiment conditions are described in Appendix A9.1. In order to investigate the effect of CO in the product on carbon formation during CO2 reforming, a feed containing CO was also tested to simulate the gas composition at 25% and 50% methane conversions in CO2 reforming. Figure 9.4 shows the amount of carbon with the time-on-stream at the following three different gas compositions: CH4 : CO2 : CO : H2 (molar ratio) = 1 : 1 : 0 : 0, 0.75 : 0.75 : 0.5 : 0.5, and 0.5 : 0.5 : 1 : 1. When the feed contained only CO2 and CH4 (molar ratio CH4 : CO2 : CO : H2 = 1 : 1 : 0 : 0), carbon was observed to form at a constant rate of 5.6 μg s−1 g−1 catalyst. However, when this molar ratio was changed to 0.75 : 0.75 : 0.5 : 0.5, the carbon formation rate increased dramatically, almost 30-fold, to 155.3 μg s−1 g−1 catalyst. Then, with even more CO present in the feed (molar ratio 0.5 : 0.5 : 1 : 1), the carbon formation rate increased even further, to 420.1 μg s−1 g−1 catalyst. Taken together, these results indicated clearly that CO in the product stream of CO2 reforming contributed greatly to carbon formation in CO2 reforming; with the possible reasons for such an increase, including: 1) The carbon species from CH4 dissociation may react readily with CO2 or surface oxygen derived from CO2 into CO, while carbon species from CO are probably less active [10]. 2.00E-03

(3) CH4:CO2:CO:H2 = 0.5:0.5:1.0:1.0 (CH4= 10 ml min–1)

1.50E-03

ATM (g)

256

(2) CH4:CO2:CO:H2 = 0.75:0.75:0.5:0.5 (CH4= 15 ml min–1) 1.00E-03

(1) CH4:CO2:CO:H2 = 1.0:1.0:0.0:0.0 (CH4= 20 ml min–1)

5.00E-04

0.00E+00 0

2000

4000

6000

8000

10000

Time (s) Figure 9.4 Carbon formation at 0.1 MPa and 923 K over 25 mg ICI R15513 catalyst at different feed gas compositions simulating (1) initial CO2 reforming CO2 : CH4 (molar ratio) = 1 : 1, 1.5 h; (2) 25% CO2 reforming conversion, 6 min; (3) 50% CO2 reforming conversion, 3 min.

9.5 TPO Analysis of Carbon Formation in CO2 Reforming

2) 1 mole of CH4 and 1 mole of CO2 can produce, stoichiometrically, 2 moles of CO. Hence, the amount of CO in the product is almost twice the amount of CH4 converted. 3) H2 is less active than CO2 in removing carbon [15]. At high CH4 conversions, less CO2 remains present.

9.5 TPO Analysis of Carbon Formation in CO2 Reforming

Based on the results of the TEOM measurements (Figure 9.4), it is possible to estimate the weight percentage of carbon formed on the ICI catalyst after a certain time on-stream. At a CH4 : CO2 : CO : H2 molar ratio of 1 : 1 : 0 : 0, a total of 4.51 wt% of carbon was formed on the catalyst after about 1.5 h. However, when the feed contained CO, much less time was required to obtain a similar amount of carbon. For example, at a CH4 : CO2 : CO : H2 molar ratio of 0.75 : 0.75 : 0.5 : 0.5, a total of 5.81 wt% carbon was obtained within approximately 6 min, whilst at a molar ratio of 0.5 : 0.5 : 1 : 1, 7.35 wt% of carbon was formed in less than 3 min (Figure 9.5). After about 1.5 h, 6 min, and 3 min of reaction during the TEOM experiments at CH4 : CO2 : CO : H2 molar ratios of 1 : 1 : 0 : 0, 0.75, 0.75 : 0.5 : 0.5, and 0.5 : 0.5 : 1 : 1, respectively, the reactants were switched to inert gas (e.g., Ar) to purge the catalyst bed and halt any further carbon formation. The catalyst samples were then removed from the TEOM sample cell, cooled to room temperature, and further assayed using a carbon analyzer (TPO-IR), in order to determine the nature of the carbon that had formed on the catalysts. As the TPO-IR is

TPO-IR Signal (a.u.)

800

(3) 600

(2)

400

(1)

200

0

273

473

673

873

1073

1273

Temperature (K) Figure 9.5 TPO-IR proﬁles of carbon formed at 0.1 MPa and 923 K over 25 mg ICI R15513 catalyst at different feed gas compositions simulating (1) initial CO2 reforming (CH4 : CO2 : CO : H2 = 1 : 1 : 0 : 0), 1.5 h;

(2) 25% CO2 reforming conversion (CH4 : CO2 : CO : H2 = 0.75 : 0.75 : 0.5 : 0.5), 6 min; (3) 50% CO2 reforming conversion (CH4 : CO2 : CO : H2 = 0.5 : 0.5 : 1 : 1), 3 min.

257

258

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

also capable of determining carbon formation quantitatively, the amounts of carbon as assessed by TPO-IR were compared with those measured with the TEOM, thus conﬁrming the ability of the TEOM for in situ carbon formation measurements. Figure 9.5 shows the TPO-IR proﬁles of three samples after TEOM experiments at CH4 : CO2 : CO : H2 molar ratios of 1 : 1 : 0 : 0 for 1.5 h, of 0.75 : 0.75 : 0.5 : 0.5 for 6 min, and of 0.5 : 0.5 : 1 : 1 for 3 min. Notably, only one distinguishable oxidation peak was apparent on all three samples. The sample after the TEOM at CH4 : CO2 : CO : H2 molar ratio = 1 : 1 : 0 : 0 showed the highest oxidation peak temperature at 868 K, while oxidation peaks occurred at 825 K and 820 K, respectively, over the other two samples after the TEOM experiment at CH4 : CO2 : CO : H2 molar ratios of 0.75 : 0.75 : 0.5 : 0.5 and 0.5 : 0.5 : 1 : 1. Chang and coworkers [16] observed only one oxidation peak at approximately 873 K in the TPO proﬁles of used supported Ni catalysts in a CO2–CH4 reaction at 973 K for several hours. However, Guola and coworkers [17] showed two oxidation peaks at 883 K and 973 K, respectively, in the TPO proﬁles over a Ni/CaO/Al2O3 catalyst after CO2 reforming at 1023 K for 5 min. Olsbye et al. [18] reported only one TPO peak at ca. 973 K over La2O3-promoted Ni/Al2O3 catalysts, but two TPO peaks at 923–973 K and >1073 K over Ni/Al2O3 after a 114 h reaction period at 973–1173 K in a CO2–CH4 reforming reaction. Shamsi and Johnson [19] observed two peaks in TPO proﬁles over a Pt/CeZrOx catalyst after CO2–CH4 reforming at 1073 K. Besides the high temperature TPO peak at ca. 973 K, a TPO peak at a temperature as low as 653 K was identiﬁed. The TPO results from Wang and Lu [20] indicated that the oxidation peak position in the TPO proﬁles also depended on the support of the catalysts. On Ni/αAl2O3, the carbon oxidation peak was at 893 K, while on Ni/MgO it was at about 923 K. Carbon formed over Ni/γ-Al2O3 and Ni/SiO2 catalysts was more difﬁcult to oxidize, with the peaks occurring around 973 K. The TPO peak temperature of 868 K on the ICI catalyst after reaction at a CH4 : CO2 : CO : H2 molar ratio o 1 : 1 : 0 : 0 for 1.5 h was generally similar to the TPO peak temperatures reported elsewhere [20]. However, the ICI catalysts after the reaction at CH4 : CO2 : CO : H2 molar ratios of 0.75 : 0.75 : 0.5 : 0.5 and 0.5 : 0.5 : 1 : 1 showed relatively lower TPO peak temperatures close to 823 K, indicating that the carbon which formed on the catalyst after reaction at CH4 : CO2 : CO : H2 = 1 : 1 : 0 : 0 had a more ordered structure. This ﬁnding was supported by the transmission electron microscopy (TEM) investigations (see the following section). Figure 9.6 compares the weight percentages of carbon estimated by the TEOM and TPO-IR on the samples after reaction at CH4 : CO2 : CO : H2 molar ratios of 1 : 1 : 0 : 0, 0.75 : 0.75 : 0.5 : 0.5, and 0.5 : 0.5 : 1 : 1. The good linear correlation between the weight of carbon measured by both methods conﬁrmed that these results were consistent. Based on this comparison, it was concluded that the TEOM was effective for studying the dynamic process of carbon formation on catalysts.

Carbon wt% measured by TPO-IR

9.6 TEM Analysis on Carbon Formed on Catalysts After CO2 Reforming of Methane 8.00 7.00

(3) 6.00

(2)

5.00 4.00

(1) 3.00 2.00 1.00 0.00 4.00

5.00

6.00

7.00

8.00

Carbon wt% measured by TEOM Figure 9.6 Comparison of the weight percentage of carbon formation measured by TEOM and TPO-IR (Carbon Analyzer) after reaction over 25 mg ICI R15513 catalysts at 0.1 MPa and 923 K at different feed gas compositions simulating (1) initial CO2

reforming (CH4 : CO2 : CO : H2 = 1 : 1 : 0 : 0), 1.5 h; (2) 25% CO2 reforming conversion (CH4 : CO2 : CO : H2 = 0.75 : 0.75 : 0.5 : 0.5), 6 min; (3) 50% CO2 reforming conversion (CH4 : CO2 : CO : H2 = 0.5 : 0.5 : 1 : 1), 3 min.

9.6 TEM Analysis on Carbon Formed on Catalysts After CO2 Reforming of Methane

The TPO-IR results obtained indicated that carbon formed at different gas compositions had different activities towards oxidation (see Figure 9.5). Consequently, to further elucidate these differences, both scanning electron microscopy (SEM) and TEM were used to analyze the carbon formed over the used ICI catalysts after a reforming reaction at 0.1 MPa and 923 K, at CH4 : CO2 : CO : H2 molar ratios of 1 : 1 : 0 : 0 for 1.5 h, of 0.75 : 0.75 : 0.5 : 0.5 for 6 min, and of 0.5 : 0.5 : 1 : 1 for 3 min. Despite a range of 4.0% to 7.0 wt% carbon being formed on the catalysts (as conﬁrmed by both TEOM and TPO-IR measurements), no differences were observed between samples in the SEM images. However, clear differences were apparent when the examination was conducted using TEM. Figure 9.7 shows the morphology of carbon on the ICI catalyst after reaction at 0.1 MPa and 923 K at a CH4 : CO2 : CO : H2 molar ratio of 1 : 1 : 0 : 0 for 1.5 h, with the amount of accumulated carbon ranging from 4.1–4.5 wt%. On examination, various types of carbon morphology were observed, which indicated that the carbon formation was a quite heterogeneous process on the catalyst, despite the average rate of carbon formation being almost constant (see Figure 9.4). Besides

259

260

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

(a)

(b)

100 nm

100 nm

Figure 9.7 Transmission electron microscopy images of the used ICI catalyst sample after 1.5 h reaction at 0.1 MPa and 923 K at feed gas composition simulating initial CO2 reforming (CH4 : CO2 : CO : H2 = 1 : 1 : 0 : 0).

encapsulating carbon, the existence of ﬁlamentous carbon with different lengths and diameters was also apparent (Figure 9.7a and b). Notably, the observed ﬁlamentous carbon had a hollow structure, with metal particles adherent to the tops of the ﬁlaments. For a catalyst sample subjected to a 6 min reaction at 0.1 MPa and 923 K at a CH4 : CO2 : CO : H2 molar ratio of 0.75 : 0.75 : 0.5 : 0.5, and for a second sample subjected to a 3 min reaction at 0.1 MPa and 923 K at a CH4 : CO2 : CO : H2 molar ratio of 0.5 : 0.5 : 1 : 1, ﬁlamentous carbon was only rarely seen, with most of the carbon being encapsulated into carbon deposits. These ﬁndings agreed well with the TPO-IR experimental results, which showed the carbon on the catalyst sample after reaction at CH4 : CO2 : CO : H2 = 1 : 1 : 0 : 0 to have the highest oxidation peak temperature among the three samples. Yet, based on all of these experimental results, it remains unclear as to whether the more ﬁlamentous carbon on the sample treated at CH4 : CO2 : CO : H2 molar ratio of 1 : 1 : 0 : 0 was due to an absence of CO in the feed, or to a longer reaction time resulting from the lower rate of carbon formation rate in the absence of CO.

9.7 Kinetic Study of Carbon Formation on PCH4 and PCO in CO2 Reforming

The TEOM results presented in Figure 9.4 showed that CO in the product stream of CO2 reforming might represent another major component contributing to carbon formation. Consequently, in an effort to further elucidate the contribution of CH4 and CO to carbon formation, a kinetics study on the effect of the partial pressure of CH4 or CO (PCH4 or PCO) on the carbon formation rate was conducted under conditions of 0.1 MPa and 923 K. Figure 9.8 shows the rate of carbon formation as a function of PCH4 or PCO. At PCH4 = 0.0062 MPa, PCO2 = 0.028 MPa, PH2 = 0.0062 MPa, and P(CO+Ar) = 0.059 MPa, the carbon formation rate was shown to increase only very slowly with the increase

9.7 Kinetic Study of Carbon Formation on PCH4 and PCO in CO2 Reforming

Carbon formation rate (μg/(s.gcat))

1200

1000

800

(3)

600

(2)

400

200

(1) 0 0

0.1

0.2

0.3

0.4

0.5

0.6

P(CH4) or P(CO) (100KPa) Figure 9.8 Carbon formation rates at different PCH4 or PCO at 0.1 MPa and 923 K over 25 mg ICI R15513 catalyst. Plot (1): P(CO+Ar) = 0.059 MPa, PCH4 = 0.0062 MPa, PCO2 = 0.028 MPa, PH2 = 0.0062 MPa.

Plot (2): P(CH4+Ar) = 0.059 MPa, PCO = 0.0062 MPa, PCO2 = 0.028 MPa, PH2 = 0.0062 MPa. Plot (3): P(Ar+CO) = 0.0375 MPa, PCH4 = 0.028 MPa, PCO2 = 0.028 MPa, PH2 = 0.0062 MPa.

of PCO (Figure 9.8, plot (1)). In fact, no carbon formation was observed until PCO exceeded PCO2 (0.028 MPa). However, at PCO = 0.0062 MPa, PCO2 = 0.028 MPa, PH2 = 0.0062 MPa, and P(CH4 + Ar) = 0.059 MPa, a much greater inﬂuence of PCH4 on carbon formation rate was evident (Figure 9.8, plot (2)) although carbon formation was only observable until PCH4 was >0.020 MPa, which was close to PCO2 (0.028 MPa). This difference in the effect of PCH4 and PCO on carbon formation most likely resulted from the inhibition of CO2 on CO dissociation over the Ni catalyst, and the promoted CH4 dissociation on Ni catalysts due to the presence of CO2 or oxygen derived from CO2 [21]. The surface carbon species from CH4 dissociation may react with surface oxygen or CO2 into other gas products such as CO, H2, and H2O, without the formation of carbon. However, when PCH4 > PCO2, the extra surface carbon species from CH4 dissociation may result in an accumulation of carbon on the catalyst surface, and lead to the observed increase in carbon formation rate. Thus, it might be speculated that the inhibitory ability of CO2 on carbon formation from CO may be weakened if more CH4 is presented in the system. Interestingly, such speculation was ﬁrmly supported by observations made during the following experiment. When both PCH4 and PCO2 were set as equal (e.g., both at 0.028 MPa), similar to the situation in the equimolar CO2–CH4 reforming reaction, PCO was observed to have the greatest effect on carbon formation rates (Figure 9.8, plot (3)). Yet, carbon

261

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

formation was observed even at PCO = 0.0 MPa. Additionally, the effect of PCO on carbon formation rates when PCH4 ≈ PCO2 was more than the sum of the observed single effect of PCH4 and PCO, as shown in Figure 9.8, plots (1) and (2). The reaction between CH4 and CO2 may weaken the inhibitory effect of CO2 on CO disproportionation and, as a result, lead to severe carbon formation. Based on the above results, it is clear that both CH4 and CO are sources of carbon formation in CO2 reforming. Indeed, CO in the products can become a major source of carbon formation in an equimolar CO2–CH4 reforming system, a ﬁnding which is consistent with the suggestion made by Bradford and Vannice [4].

9.8 H2O Effect on Carbon Formation in CO2 Reforming

By conducting a similar kinetic study, the changes in carbon formation rate were investigated by replacing part of the CO2 with H2O. As shown in Figure 9.9, the presence of H2O can signiﬁcantly suppress the carbon formation rates from both CH4 and CO. For example (Figure 9.9, plot (3)), when comparing the effect of PCO on carbon formation in the case of PCH4 = 0.028 MPa and PCO2 = 0.028 MPa, and in the case of PCH4 = 0.028 MPa and PCO2 = PH2O = 0.014 MPa, the carbon formation rates increased sharply when PCO increased from 0 to 0.03 MPa in the former case. In contrast, in the latter case there was barely any carbon formation until

Carbon formation rate (μg s–1 g–1 cat)

262

1200.00

1000.00

800.00

600.00

400.00

200.00

(2) (1)

(3) 0.00 0

0.1

0.2

0.3

0.4

0.5

0.6

P(CH4) or P(CO) (100KPa) Figure 9.9 Carbon formation rates at different PCH4 or PCO at 0.1 MPa and 923 K over 25 mg ICI R15513 catalyst. Plot (1): P(CH4+Ar) = 0.059 MPa, PCO2 = PH2O = 0.014 MPa, PH2 = PCO = 0.0062 MPa. Plot (2): P(CO+Ar) =

0.056 MPa, PCH4 = 0.012 MPa, PH2O = PCO2 = 0.013 MPa, PH2 = 0.0062 MPa. Plot (3): P(CO+Ar) = 0.038 MPa, PCH4 = 0.028 MPa, PH2O = PCO2 = 0.014 MPa, PH2 = 0.0062 MPa.

References

PCO = 0.03 MPa, even though a slight increase in carbon formation rate was observed when PCO was >0.03 MPa. Thus, the addition of H2O into the CO2 reforming system can signiﬁcantly reduce carbon formation, this being most likely due to the stronger ability of H2O to remove carbon species from both CH4 and CO.

9.9 Conclusions

Kinetics studies on carbon formation, using the TEOM, have established that both CH4 in the reactants and CO in the products, may serve as the source of carbon formation in the CO2 reforming reaction. In an equimolar CO2–CH4 reforming condition, CO in the product stream is most likely the major source of carbon formation. However, carbon formed in CO2 reforming from a feed without containing CO shows a more ﬁlamentous morphology, which is more difﬁcult to be oxidized; this contrasts with carbon formed from a feed containing CO, which mostly encapsulates metal particles in the catalyst and is relatively easily oxidized. Unfortunately, it is currently unclear as to whether these differences result from the presence of CO in the feed, or from the different times on-stream to accumulate similar amounts of carbon due to different carbon formation rates when the feeds contain CO or no CO. The replacement of some of the CO2 with H2O may greatly inhibit the carbon formation encountered in CO2 reforming, with kinetics studies having shown that H2O can reduce carbon formation from both CO and CH4.

Acknowledgments

The authors wish to acknowledge the US Department of Energy, National Energy Technology Laboratory for the partial ﬁnancial support of this work. They are also very grateful for the helpful suggestions made by Dr M.A. Vannice and Dr S. Eser of PSU.

References 1 (a) Aresta, M. (ed.) (2003) Carbon Dioxide Recovery and Utilization, Springer, New York, p. 407. (b) Song, C.S., Gaffney, A.M., and Fujimoto, K. (eds) (2003) CO2 Conversion and Utilization, ACS Symposium Series, American Chemical Society Publication, Washington, DC, p. 440. (c) Liu, C.-J., Mallinson, R.G., and

Aresta, M. (eds) (2003) Utilization of Greenhouse Gases, ACS Symposium Series, American Chemical Society Publication, Washington, DC, p. 424. (d) Song, C.S. (2006) Catal. Today, 115, 2–32. (e) Ashcroft, A.T., Cheetham, A.K., Green, M.L.H., and Vernon, P.D.F. (1991) Nature, 352 (6332), 225–226.

263

264

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

2 3

4 5 6 7

8

9

10

11

12

(f) Song, C.S. and Pan, W. (2004) Catal. Today, 98, 463–484. (g) Song, C.S. (2001) Chem. Innov., 31, 21–26. Wang, S. and Lu, G.Q. (1998) Energy Fuels, 12 (6), 1235–1240. Tomishige, K., Chen, Y.G., and Fujimoto, K. (1999) J. Catal., 181, 91–103. Bradford, M.C.J. and Vannice, M.A. (1996) Appl. Catal. A, 142 (1), 73–96. Richardson, J.Y. and Paripatyadar, S.A. (1990) Appl. Catal., 61 (2), 293–309. Zhang, Z.L. and Verykios, X.E. (1994) Catal. Today, 21 (2/3), 589–595. Baker, R.T.K. and Harris, P.S. (1978) Chemistry and Physics of Carbon, vol. 14 (eds P.L. Walker and P.A. Thrower), Marcel Dekker, New York, pp. 83–165. Tomishige, K., Himeno, Y., Matsuo, Y., Yoshinaga, Y., and Fujimoto, K. (2000) Ind. Eng. Chem. Res., 39, 1891–1897. Swaan, H.M., Kroll, V.C.H., Martin, G.A., and Mirodatos, C. (1994) Catal. Today, 21 (2/3), 571–578. Olsbye, U., Wurzel, T., and Mleczko, L. (1997) Ind. Eng. Chem. Res., 36 (12), 5180–5188. Nolan, P.E., Lynch, D.C., and Cutler, A.H. (1998) J. Phys. Chem. B, 102, 4165–4175. Dent, F.J., Moignard, L.A., Eastwood, A.H., Blackburn, W.H., and Heeden, D. (1945–1946) An Investigation into the Catalytic Synthesis of Methane, Gas Research Board, London, pp. 604–693.

13 Rostrup-Nielsen, J.R. (1972) J. Catal., 27, 343–356. 14 Denbigh, K. (1971) The Principles of Chemical Equilibrium, 3rd edn, Cambridge University Press, Cambridge, Great Britain, p. 494. 15 Bridger, G.W., and Chinchen, G.C. (1970) Hydrocarbon-reforming catalysts, in Catalyst Handbook – with Special Reference to Unit Processes in Ammonia and Hydrogen Manufacture, SpringerVerlag, New York, Ch 5, pp. 63–96. 16 Chang, J.S., Park, S.E., and Chon, H. (1996) Appl. Catal. A: General, 145 (1/2), 111–124. 17 Goula, M.A., Lemonidou, A.A., and Efstathiou, A.M. (1996) J. Catal., 161, 626–640. 18 Olsbye, U., Moen, O., Slagtern, A., and Dahl, I.M. (2002) Appl. Catal. A: General, 228 (1), 289–303. 19 Shamsi, A. and Johnson, C.D. (2001) Am. Chem. Soc., Div. Fuel Chem. Prepr., 221, 49. 20 Wang, S. and Lu, G.Q.M. (1998) Appl. Catal. B: Environmental, 16 (3), 269– 277. 21 Luo, J.Z., Gao, L.Z., Ng, C.F., and Au, C.T. (1999) Catal. Lett., 62 (2/4), 153–158. 22 Demicheli, M.C., Ponzi, E.N., Ferretti, O.A., and Yeramian, A.A. (1991) Chem. Eng. J. Biochem. Eng., 46 (3), 129–136. 23 Kroll, V.C.H., Swaan, H.M., and Mirodatos, C. (1996) J. Catal., 161 (1), 409–422.

Appendix A9.1 A Brief Description of the TEOM, TPO-IR, and TEM Experiments TEOM Measurements The TEOM (Rupprecht & Patashnick, Co., Inc.) measures the weight change of a sample in situ by comparing the vibrating frequency of the tapered glass sample cell during the measurement. At the start of an analysis, the sample cell vibrates in a set frequency, depending on its mass. When there is mass change of the sample, the vibration frequency of the sample cell consequently changes. The frequency change of the sample cell is detected by an optical device and converted into the mass change, based on Equation A1.

Appendix A9.1

Δm = k f 12 − k f 02

(A1)

where k = constant for a speciﬁc TEOM apparatus, and f0 and f1 are the frequencies of the glass sample cell at the measurement times of t0 and t1, respectively. A catalyst sample was loaded into the tip of the tapered glass cell (reactor) and supported and ﬁxed by quartz wool and a metal cap coated with gold. The glass cell was protected by a stainless-steel tube. The temperature of the TEOM was controlled by the temperature control unit and software, while the pressures in the system and the ﬂows of input gases were controlled by a back-pressure regulator (TESCOM) and mass ﬂow controllers (Brooks), respectively. Distilled water was pumped into the system using an ISCO syringe pump (Model 500D). A purge gas (Ar, 100 ml min−1) was used to sweep outside the glass cell in order to carry the efﬂuent from the reactor out of the system and, in the meantime, to prevent efﬂuent from ﬂowing into the optical devices. One of the major advantages of the TEOM compared to conventional gravimetric microbalance systems is that the conﬁguration of the TEOM enables all reactants to pass through the catalyst bed inside the glass sample cell (as in a ﬁxed-bed reactor), thus avoiding the problems of gas bypass and buoyancy which are often encountered in a conventional thermogravimetric analysis (TGA) measurement. Demicheli et al. [22] noted that a high ﬂow rate of feed could affect the weight variation measurements when using a TGA apparatus; hence, the feed gas ﬂow rates were limited to a very low range. Kroll et al. [23] observed a large discrepancy of carbon deposition measured by the TGA and by TPH/TPO over catalyst samples after different times on-stream in a reactor. This discrepancy was attributed to the higher temperature gradients throughout the catalyst bed in the TGA due to the gas bypass. The microbalance used by Kroll et al. was equipped with a perforated basket instead of a plug-ﬂow, ﬁxed-bed reactor. The in situ measurement of carbon deposition by the TEOM provides more accurate information on carbon formation behavior than do ex situ approaches such as TPH and TPO, especially when the carbon formation rates change with the time on-stream. The TEOM experiments described in this chapter were carried out at 923 K and 0.1 MPa. Prior to each measurement, 12–50 mg ICI catalyst (35–60 mesh) was loaded into the glass tube and reduced by 40% H2 in Ar (5 ml min−1 H2 + 7.5 ml min−1 Ar) at 373 K for 10 min, 723 K for 75 min, and 923 K for 30 min. The heating rate from room temperature to 373 K, from 373 K to 723 K, and from 723 K to 923 K was 12 K min−1. After the reduction, the catalyst was purged with Ar (7.5 ml min−1) for 10 min, followed by the preparation of reactant gases in another line. After another 5 min, the prepared gases were switched into the TEOM and the monitoring of mass changes started. When the mass change was found to exceed 15 wt% of the sample, the experiment was stopped to protect the glass cell from damage. In the case when H2O addition was necessary, H2O was pumped into the TEOM through a separate line connected to the preheating zone of the TEOM. The gases and vaporized steam in the preheating zone were mixed and ﬂowed into the TEOM.

265

266

9 In-Situ Study of Carbon Deposition during CO2 Reforming of Methane

Carbon Analyzer The amount of carbon deposited on the catalysts after TEOM experiments was determined using a carbon analyzer (LECO RC-412). Samples of catalyst (10– 20 mg) after reforming reactions were loaded into the sample holder and held in the heating zone, purged by O2 (750 ml min−1). After starting the analysis, the temperature of the heating zone was increased from 373 K to 973 K at a rate of 30 K min−1. When the heating zone temperature reached a certain high temperature, which depended on the nature of carbon in the sample, the carbon on the sample could be oxidized into CO2. The CO2 produced was monitored using an IR detector, and the amount of carbon on the sample was calibrated by comparison with a standard sample (LECO calibration sample with 1.03% carbon content). The temperature at which the carbon was oxidized provided information on the activity of carbon on the sample towards oxidation. Transmission Electron Microscopy The morphology of carbon on the Ni/Al2O3 catalyst (ICI catalyst) after CO2 reforming was determined with TEM (point-to-point resolution of 1 nm) using an ultrasonically dispersed (in ethanol) catalyst sample deposited on a thin carbon ﬁlm supported on a standard copper grid. The microscope (JEOL JEM 1200 EXII) was interfaced with a video camera and a high-resolution Tietz F224 camera. The accelerating voltage during the scanning was 80 kV.

267

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches Ji-Jun Zou and Chang-Jun Liu

10.1 Introduction

The consumption of all fossil fuels leads to the formation of carbon dioxide (CO2) which, by itself has little value by far, yet it contributes approximately 57% to manmade “greenhouse gases” [1]. Any success in the research and development of a feasible CO2 utilization would signify the achievement of two objectives: (i) to slow down the build-up of greenhouse gases in the atmosphere; and (ii) to provide a better carbon resource utilization. Unfortunately, as the CO2 molecule has very low energy content, a large amount of additional energy (that, of course, would in turn induce more CO2 emissions!) or expensive hydrogen is required for the conventional catalytic hydrogenation of CO2. If CO2 were to replace other carbon sources, then its use cannot produce more CO2 than can be stored. Consequently, those reactions based on CO2 must have reaction energies less than those of the substituted reactions, as well as a better yield and selectivity. This could be achieved by using a more direct synthetic pathway compared to existing methodologies. Aresta and Dibenedetto [2] categorized the ﬁve main reactions that may contribute to reducing the atmospheric loading, while generating a proﬁt, as: 1) Direct carboxylation reactions (generation of the moieties C–COOH, C–COOR, N–COOR, N–CO–N, –NCO). 2) The use of CO2 as oxidant in selective processes. 3) The use of CO2 as an additive to CO for the synthesis of methanol. 4) Direct on-site liqueﬁed natural gas to liquid (GTL) conversion. 5) The use of supercritical CO2 (scCO2) as a solvent, or as a solvent and reagent. Recently, unusual plasma chemistry (especially nonthermal plasma) has attracted much attention with regards to the effective and efﬁcient activation of CO2. These nonthermal plasma approaches can even be performed at room temperature and atmospheric pressure. As nonthermal plasmas have also been successfully applied for the removal of NOx and SO2 from ﬂue gases, an increasing number of investigations has been conducted to examine their use for CO2 utilization. Nonthermal plasma approaches for CO2 utilization were ﬁrst Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

268

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

summarized in 1999 [3], since which time excellent progress has been made in the ﬁve categories denoted by Aresta and Dibenedetto [2]. The aim of this chapter is to provide a comprehensive summary of the present art of CO2 utilization through nonthermal plasmas.

10.2 Nonthermal Plasma Phenomena

In general, plasma is an ionized gas that can be generated by several methods that include combustion, ﬂame, electrically heated furnaces, electric discharge (corona, spark, glow, arc, microwave discharge, plasma jets and radiofrequency plasma), and shocks (whether electrically, magnetically, and chemically driven) [4]. Depending on their energy level, temperature and ionic density, plasmas are usually classiﬁed as high-temperature plasma, thermal plasma and nonthermal or nonequilibrium plasma. Among these, nonthermal plasma has been applied to ﬂue gas treatment, and has in the past been considered to show much promise for organic syntheses on the basis of its nonequilibrium properties, low power requirements, and its capacity to induce physical and chemical reactions within gases at relatively low temperatures. The electrons in nonthermal plasma can reach temperatures of 10 000–100 000 K (1–10 eV), while the gas temperature can remain as low as room temperature. It is the high electron temperature that determines the unusual chemistry of nonthermal plasmas. Based on the mechanisms by which plasma is generated, the pressure applied and the electrode geometry, nonthermal plasmas can exist as several very different types, including glow discharge, corona discharge, silent discharge or dielectric-barrier discharge, microwave discharge and radiofrequency (RF) discharge [4–7]: Glow discharge: This is a low-pressure discharge (<1 kPa) that usually operates between ﬂat electrodes. The electrons in the glow discharge are highly energetic, and the excited neutral atoms and molecules generate a typical glow (e.g., ﬂuorescent tubes). Due to its low-pressure characteristics, the glow discharge is not really suitable for chemical synthesis. Corona discharge: This is an inhomogeneous discharge, and can be initiated at atmospheric pressure using inhomogeneous electrode geometries, such as a pointed wire electrode with a plate geometry. It is the small radius of curvature at the top of the wire electrode that results in the high electric ﬁeld required to ionize the neutral molecules. Initially, the corona discharge was considered to be the most effective technique for producing CO from CO2 [8]. Silent discharge or dielectric-barrier discharge (DBD): This combines the largevolume excitation of the glow discharge with the high-pressure characteristics of the corona discharge. In the silent discharge a dielectric layer covers at least one electrode, such that the entire electrode area will be effective for discharge reactions. When the silent discharge is initiated at any location within the gap between electrodes, charge accumulates on the dielectric to form an opposite

10.2 Nonthermal Plasma Phenomena

electric ﬁeld and interrupt the current ﬂow in a few nanoseconds to generate microdischarges. The duration of the current pulse relates to the pressure, the properties of gases and the dielectric material applied. RF discharge: This operates at high frequencies (several MHz) and very low pressures to achieve nonequilibrium conditions. This discharge is also not suitable for chemical synthesis. Microwave discharge: This operates at very high frequencies (e.g., 2.45 GHz) in the range of microwaves, within which only light electrons can follow the oscillations of the electric ﬁeld. Therefore, this discharge is far from the local thermodynamic equilibrium, and can be operated over a wide pressure range. The performance of a microwave discharge depends heavily on the type of microwave power applicator (detailed information on this subject is available elsewhere [6, 7]). Gliding arc discharge: This is a combination of high-power equilibrium arc discharge and better selectivity of nonthermal plasmas. It has also been reported to be used for the utilization of greenhouse gases such as CH4 and/or CO2 [9–13]. A comparison of the electron energy and gas temperatures of some gas discharges is provided in Table 10.1, while the energy parameters of gases involved in plasma CO2 utilization are summarized in Table 10.2. The fundamental respects Table 10.1

Characteristic parameters of gas discharge [5].

Pressure Electric ﬁeld Reduced ﬁeld Electron energy Electron density Degree of ionization

Table 10.2

Glow discharge

Corona discharge

Silent discharge

<1 kPa 10 V cm−1 50 Td 0.2–2 eV or 5000–20 000 K 108–1011 cm−3 10−6–10−5

100 kPa 0.5–10 kV cm−1, variable 2–200 Td, variable 5 eV, variable

100 kPa 0.1–100 kV cm−1 1–500 Td 1–10 eV

103 cm−3, variable small, variable

1014 cm−3 10−4

Energy parameters of gases.

Gas

Dissociation energy (eV)

Ionization energy (eV)

CO2 CO H2 O2 H2O C2H4 C3H8 C2H2 CH4

5.5 11.1 4.5 5.1 – – – – −10

14.3 14.0 15.4 12.5 12.8 10.5 11.2 11.4 13.0

269

270

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

of nonequilibrium plasma phenomena have been discussed elsewhere [4–7, 9], and so will not be described at this point. As nonthermal plasma is a mixture of electrons, highly excited atoms and molecules, ions, radicals, photons, and so on, its chemistry is extremely complex, and highly selective products should not be expected via plasma chemistry. The basic reactions for controlling both the direction and reaction rate of plasma CO2 utilization can be summarized as follows (here, A and B represent atoms, A2 and B2 molecules, e represents an electron, M is a temporary collision partner, and S represents a solid surface site. The excited species is indicated by an asterisk). 10.2.1 Electron/Molecular Reactions

•

Excitation (rotational, vibrational, electronic): e + A 2 → A 2* + e

•

Dissociation: e + A2 → 2 A + e

•

(10.7)

Detachment: e + A 2− → A 2 + 2 e

•

(10.6)

Recombination: e + A 2+ → A 2

•

(10.5)

Dissociation ionization: e + A2 → A + + A − + 2 e

•

(10.4)

Ionization: e + A 2 → A 2+ + 2 e

•

(10.3)

Dissociation attachment: e + A2 → A − + A

•

(10.2)

Attachment: e + A 2 → A 2−

•

(10.1)

(10.8)

Decomposition: e + AB → A + B + e

(10.9)

10.2.2 Atom/Ion/Molecule Reactions

The reactions among atoms, ions and molecules are mostly related to inelastic collisions between heavy particles, including:

10.2 Nonthermal Plasma Phenomena

•

Penning dissociation: M* + A 2 → 2 A + M

•

Penning ionization: M* + A 2 → A 2+ + M + e

•

(10.15)

Neutral recombination: A + B + M → AB + M

•

(10.14)

Associative attachment: A − + A → A2 + e

•

(10.13)

Collisional detachment: M + A 2− → A 2 + M + e

•

(10.12)

Ion recombination: A + + B− → AB

•

(10.11)

Charge transfer: A ± + B → B± + A

•

(10.10)

(10.16)

Synthesis: A + B → AB

(10.17)

A* + B → AB

(10.18)

10.2.3 Heterogeneous Reactions

•

Neutral recombination: S-A + A → S + A 2

•

Metastable de-excitation: S + M* → S + M + hv

•

(10.20)

Neutral abstraction: S− B + A → S + AB

•

(10.19)

(10.21)

Sputtering: S− B + M* → S+ + B + M

(10.22)

In addition to these reactions, others (e.g., photochemical reactions) also occur in the plasmas (such details are not included at this point). The rates of these discharge reactions depend on the electron energy, electron density, gas temperature, gas pressure, and properties of the gases, and so on. However, as the present

271

272

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

understanding of plasma chemistry is limited, it remains difﬁcult to predict ﬁnal products on a theoretical basis.

10.3 CO and/or H2 Production from CO2 10.3.1 CO2 Dissociation

The dissociation of CO2 to CO is the simplest reaction for CO2 reduction, and is also the primary reaction for all CO2 utilization reactions: CO2 → CO + 1 2 O2, ΔG1000K = 190.5 kJmol −1

(10.23)

It is clear that, even for this reaction, there exist certain difﬁculties in the thermodynamics, because ΔG increases by the reaction going from left to right. According to Venugopalan and Veprek [4], plasma technology is attractive under two scenarios: (i) that reactions are possible but kinetically hindered without a plasma, such as CO2 dissociation; and (ii) that reactions are impossible due to an extreme chemical equilibrium, as in the reaction: CO2 → C + O2, ΔG1000K = 2826.1 kJmol −1

(10.24)

For scenario (i), a “weak” or nonthermal plasma is sufﬁcient, but for scenario (ii) a high-energy or thermal plasma is required. Many important reactions for CO2 utilization meet the requirement of scenario (i), which occurs mostly via the dissociation of CO2: CO2 + e → CO2* + e → CO + O + e

(10.25)

→ CO + O

(10.26)

This is a basic reaction for all CO2-involved plasma reactions. Some primary investigations have been conducted for the CO2 dissociation using corona discharge [8, 14–17]. Some types of CO2 adsorbent, such as basic zeolite, can result in a much higher CO2 dissociation rate under corona discharge [18]. Suib et al. [19–21] studied the effects of coatings on the rod and diluent gas in a fan-type ac glow discharge plasma reactor. Here, the diluent gas plays an important role in the activation of CO2, via charge and energy transfer from excited state dilute gas species to produce vibrationally excited CO2+ intermediates. As evidence, the order of reactivity depends on the rate constants for bimolecular energy transfer to CO2: 16 × 10−10 cm3 s−1 (He2+ ); 11 × 10−10 cm3 s−1 ( Ar2+ ); and 7.7 × 10−10 cm3 s−1 ( N2+ ). The order of reactivity for the different metal catalyst coatings was Cu > Au > Rh > Fe ≈ Pt ≈ Pd. Although the effect of the metal electrode seemed very complex, there was a clear relationship between the excitation temperature of pure He plasma and the discharge metals, with a high frequency, a high ﬂow

10.3 CO and/or H2 Production from CO2

rate, and a relatively high CO2 concentration being preferred for high reaction rates and high energy efﬁciencies. When Andreev et al. [22] investigated a non-selfsustained glow discharge, it was found that the energy costs could be reduced more than an order of magnitude by means of controlling the electronic component such as the E/N (E is the longitudinal electrical ﬁeld strength and N is the density of neutral plasma gas particles). A dielectric barrier discharge plasma has also been used for the decomposition of CO2. The efﬁciency of the DBD reaction should be increased by increasing the permittivity of the barrier material, although a high permittivity would tend to be broken if a high voltage were to be supplied, because of its modest dielectric strength. Therefore, materials with high dielectric constant and high dielectric strength are desirable as a dielectric barrier. Li et al. [23–25] synthesized a new dielectric barrier of CaxSr1–xTiO3 (x < 0.4) by combining the high permittivity of SrTiO3 with the high dielectric strength of CaTiO3. The mechanical and dielectric properties were further enhanced by adding Li2Si2O5. At 373 K and 10 MHz, the order of permittivity was Ca0:7Sr0:3TiO3 (207) > alumina (10.4) > silica glass, and the CO2 conversion was proportional to the permittivity (15% > 6% > 4%, respectively). Later, Indarto et al. [26] reported that a gliding arc plasma had a higher power efﬁciency than other nonthermal plasma methods. Although the conversion percentage was relatively similar compared to the DBD discharge, this system was shown capable of handling a higher input ﬂow rate (∼40-fold higher). 10.3.2 Reforming of CH4 with CO2

The higher hydrocarbon formation from syngas has long been industrialized as the Fischer–Tropsch synthesis [27]. Yet, syngas production from the CO2 reforming of methane is an endothermic reaction, and requires a high temperature (ca. 1073 K) for a favorable equilibrium: CH4 + CO2 → 2 CO + 2 H2, ΔH ° = 258.9 kJmol −1

(10.27)

Another drawback of the CO2 reforming of methane is that, with the present catalyst designs, undesirable carbon deposits are unavoidable. Nonthermal plasmas have also been considered as alternatives for syngas production from CH4 and CO2. Both, Gesser et al. [28] and Kogelschatz et al. [29, 30] have reported syngas formation from CO2 reforming of methane using the silent discharge. Motret et al. [31] measured the rotational temperature and reactor wall temperature of CH4 + CO2 plasma in a silent discharge at atmospheric pressure, by using spectroscopy. In the reaction, the gas temperature inside the streamers was approximately 3000 K, while the mean gas temperature in the active volume was close to room temperature (<350 K). This indicated that, by using this process, syngas formation could be carried out at ambient temperatures, with a sufﬁciently high yield.

273

274

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

Gesser et al. [32] suggested a very different mechanism from that proposed for the catalytic CO2 reforming of methane, that involved the reactions of ions and radicals. However, based on the energy parameters listed in Tables 10.1 and 10.2, the mechanisms would be mostly dominated by radical reactions. Kogelschatz et al. [29] also suggested a radical reaction mechanism in which the dissociation of CO2 and CH4 molecules was suggested as being responsible for the initiation of radical reactions: e + CH4 → CH3 + H + e

(10.28)

e + CO2 → CO + O + e

(10.29)

Methyl radicals produced in this way would also induce a higher hydrocarbon and oxygenates formation in some cases (this will be discussed later). The H2/CO ratio in syngas produced using a silent discharge depends on the feed condition. In fact, it was noted that the syngas yield increased almost linearly with an increasing discharge power, whereas a higher discharge power led to a higher energy efﬁciency [28–30]. However, the energy efﬁciency was still too low in both cases. The highest energy efﬁciency (deﬁned as the electric energy converted to chemical energy in syngas) was reported to be approximately 7% [29]. In an effort to improve the energy efﬁciency, further investigations have been conducted, including the use of catalysts and an optimization of the discharge parameters. Song et al. [33] reported that a Ni/γ-Al2O3 catalyst could increase the selectivity of CO by inhibiting the formation of C2–C4 hydrocarbons. A downstream thermal catalytic zone was further applied to enhance the process [34]. With an input power of 80 W in the plasma zone and a temperature of 573 K in the downstream catalytic zone, the conversion of CH4 was 97.4%, and that of CO2 almost 100%. Notably, the conversion of CH4 was 40% using DBD, and only 2% using a catalyst. The selectivity of H2 was almost 100% and the selectivity of CO was 96.9%. This system proved to be very effective, not only in terms of reaction performance but also in its economy. The same group also compared the performance of bipolar and unipolar pulse power supplies [35] showing that, for the same amount of energy consumed, the conversion in bipolar pulse was almost doubled compared to a unipolar pulse. In addition, Li et al. [36] found that the electrode materials could have a signiﬁcant effect on the reaction activity. For the feed of methane, the order of conversion was Ti ≈ Al > Fe > Cu, while the order was Al > Cu > Ti > Fe in the case of a CO2 feed without methane. However, when methane and CO2 were fed simultaneously, the Ti electrode showed the best activity for the conversions, while the other three materials showed similar performances; however, the metals had similar inﬂuences on the distribution of the product. Sarmiento et al. [37] reported that a controlled roughness of the surface would be beneﬁcial in sustaining the silent discharge. Corona discharges have also attracted much attention for the CH4/CO2 reforming reactions. Li et al. [38–40] conducted the reaction via AC and DC corona discharge plasmas at atmospheric pressure. The H2/CO ratio was seen to depend heavily on the CH4/CO2 ratio in the feed, increasing from 0.21 at 0.2 CH4/CO2 to

10.3 CO and/or H2 Production from CO2

2.15 at 2.0 CH4/CO2. The conversions of CH4 and CO2 were increased as the discharge power was raised, and were decreased as the ﬂow rate was increased. The conversions obtained in the reactor below 380 K were higher than the equilibrium conversions at 890 K, thus demonstrating the advantages of plasma chemistry compared to thermal chemistry. The conversions followed the order: positive corona > AC corona > negative corona, whereas H2/CO ratios in the products exhibited the opposite order. When Yao et al. [41, 42] used a high pulse frequency plasma to ignite the corona plasma, the conversions of CH4 and CO2 and the energy efﬁciency were each improved; however, the selectivity of CO was decreased because a considerable amount of C2 hydrocarbons had been formed. (This combined process of reforming and oxidative coupling of methane will be discussed later in the chapter.) Later, Suib et al. [43] designed a Y-type ac discharge reactor with two arms, into which CO2 and CH4 could be introduced and activated separately. By using this approach, interactions between species from different arms could be observed, yet the interaction was also observed when only one reactant was excited. Micro-arc formation between excited CH4 and excited CO2 was shown to increase the conversion of both reactants and to favor the production of CO. However, these plasma approaches have much lower energy efﬁciencies compared to conventional catalytic methods. Zhang et al. [44] utilized a high voltage, wire-like plasma that was enhanced and spread by a pulsed microwave for the oxidative coupling and reforming of CH4 with CO2. The energy efﬁciency was clearly improved when compared to dielectricbarrier discharges, high-frequency pulsed plasma, and other microwave plasmas. Both, Lesueur et al. [45] and Czernichowski [46] achieved very high energy efﬁciencies (close to 40%) when using the gliding arc discharge. The highlight of the gliding arc discharge method was the combination of a high-power thermal plasma and a better selectivity of nonthermal plasma. In fact, up to 3 kW electric power was applied [45], which was much higher than that used for the silent discharge [28–30]. An additional report on the nonequilibrium atmospheric pressure gliding arc discharge for CO2 reforming was presented by Kennedy et al. [10], in which the energy efﬁciency was drastically improved by utilizing the energy of the exhausted combustion gases. Indarto [47] found C2H2 also to be a major product when using gliding arc discharge, the selectivity of which was as high as 40%. When Ghorbanzadeh et al. [48] applied an atmospheric pulsed glow discharge, sustained by corona pre-ionization, to achieve a uniform glow discharge in a large volume at atmosphere pressure, the reforming efﬁciency was comparable to that obtained with gliding arc plasmas, but the system had the advantage that it could be operated at near room temperature. 10.3.3 Reforming of Aliphatic Hydrocarbons with CO2

Similar to the reforming of methane, the major product of aliphatic hydrocarbons reforming is syngas, although some dehydrogenated products such as oleﬁns and alkynes may also be formed. Futamura et al. [49–51] reported the CO2 reforming

275

276

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

of methane, propane and neopentane using a silent discharge. In this case, at 433 K, the energy efﬁciencies for H2 formation were 0.012, 0.048, and 0.037 g H2 kWh−1, respectively, which was rationalized in terms of the bond dissociation energies of C–H: [H−CH(CH3)2] (399.6 kJ mol−1) < [H−CH2C(CH3)3] (418.8 kJ mol−1) < (H−CH3) (438.9 kJ mol−1). Propane is the most promising candidate to provide higher H2 yields in CO2 reforming. In this case, CO was formed not only from CO2 but also from the oxidation of hydrocarbons, such that a higher CO2/C3H8 ratio would be needed to oxidize C3H8 carbon atoms to CO. The positive temperature effect on the reaction could be ascribed to the promotion of a secondary decomposition of the hydrocarbons, induced by radicals formed in situ. When Zhang et al. [52] studied the dehydrogenation of ethane with CO2 under a pulse corona plasma, the conversion of ethane increased monotonously with the increase in CO2 concentration in the feed. The maximum yield of ethylene and acetylene appeared with the fraction of CO2 as 50%. The use of CeO2/γ-Al2O3 and La2O3/γ-Al2O3 catalysts increased the conversion of C2H6, and also the selectivity and yield of C2H4 and C2H2, but the conversion of CO2 was slightly reduced. The selectivity of C2H4 was shown to increase signiﬁcantly when Pd/γ-Al2O3 was used as the catalyst. 10.3.4 Other Reforming Reactions with CO2

Futamura et al. [53] reported the reforming of CO2 using H2O to produce CO and H2 in a silent discharge plasma reactor: CO2 + H2O → CO + H2

(10.30)

An arbitrary H2/CO ratio was obtained by controlling the ration of H2O to CO2., while the reaction was also improved by packing ferroelectric into the discharge gap. The energy conversion efﬁciency was shown to decrease with the water content. CO2 reforming with H2S using microwave discharge was also reported to produce syngas. The products contained the elements sulfur and syngas, with a 1/1 ratio of H2/CO: CO2 + H2S → CO + H2 + H2O + 2 S

(10.31)

Potapkin et al. [54] showed this reaction to be a two-stage process, whereby H2S was ﬁrst dissociated plasma-chemically and hydrogen produced. The hydrogen could then react with CO2 to generate water and CO. The overall effect of CO2 was to increase the conversion of H2S, but to reduce the hydrogen yield. 10.3.5 Reduction of CO2

Nakagawa et al. [55] reported their investigations into the reduction of CO2 using a pulse power discharge, a discharge phenomenon which was similar to the pulse-

10.4 Hydrocarbons Synthesis from CO2

corona-induced plasma chemical processing, in the presence/absence of hydrogen. CO2 reduction in the presence of hydrogen was found to be independent of the gas pressure. For example, at low pressure (<60 kPa) the reduction efﬁciency was almost doubled when hydrogen was present in the CO2 feed, whereas at 101 kPa the CO2 reduction was similar whether hydrogen was present, or not. These authors suggested that an arc discharge could occur at pressures below 60 kPa, but the effect could also be related to the higher electron energy at lower pressures. The reverse water gas shift reaction would be more favored under such conditions: CO2 + H2 → CO + H2O, ΔG500K = 20.4 kJmol −1

(10.32)

10.4 Hydrocarbons Synthesis from CO2 10.4.1 Oxidative Coupling of CH4 with CO2

Based on the ﬁndings of studies in catalysis, CO2 may represent a better oxidant than oxygen for the oxidative conversion of methane to higher hydrocarbons. The most commonly considered products for such methane conversion are the C2 hydrocarbons: CO2 + 2 CH4 → CO + C2H6 + H2O, ΔG1073K = 35 kJmol −1

(10.33)

CO2 + C2H6 → CO + C2H4 + H2O, ΔG1073K ≈ 0 kJmol −1

(10.34)

and

At 1073 K, the equilibrium yields of ethane and ethylene are fairly high (13% and 57%, respectively) [56]. However, the experimental yield of C2 hydrocarbons over metal oxide catalysts was not sufﬁciently high; in fact, when compared to CO2 reforming of methane the oxidative coupling reactions were less-favored thermodynamically. Gas discharges might offer a good alternative to these conventional catalytic conversions, as discussed previously. For example, Liu et al. [57, 58] conducted an initial investigation into the oxidative conversion of CH4 using CO2 as an oxidant via a streamer AC corona discharge. As a consequence, a higher C2 yield (ca. 35%) was achieved, and the maximum energy yield of C2 hydrocarbons was 5.0 g kWh−1. In this case, the major product was CO, the maximum yield of which was 29.7 g kWh−1. Yao et al. [41] showed that, with a coaxial cylindrical reactor, the selectivity of C2H6 based on carbon atoms of CH4 was 60% at 773 K with the pulse frequency of 500 pulses per second (PPS), whereas the selectivity of C2H4 was 64% with a frequency of 2920 PPS. It seemed that a high pulse frequency favored the formation of ethylene. The energy efﬁciency could be improved by using a high temperature and a high pulse frequency. In a point-to-point reactor without external heating [42], a high selectivity of C2H2 (up to 11.3%) was achieved, together

277

278

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

with a (CH4 + CO2) conversion efﬁciency of 2.4 mmol kJ−1. The C2H2 selectivity could be moderated by changing the CH4 concentration in the feed mixture. Zhang et al. [44] also reported the formation of a coupled product in the CH4/CO2 reforming by using pulsed microwave plasma. Here, under a condition of 1.5 : 1 volume ratio of CH4 to CO2, a total ﬂow rate of 200 ml min−1, a microwave peak power of 120 W, and a pulse duty factor of 100 ms/100 ms, the conversions of CH4 and CO2 were approximately 70% and 68%, and the selectivities of C2H2 and C2H4 17.8% and 4.1%, respectively. Notably, the coproduction of acetylene of high value remarkably improved the economical perspective of syngas production. 10.4.2 Hydrogenation of CO2/CO to C2 Hydrocarbons

Scientists in the former Soviet Union have developed a series-connected discharge catalytic CO2 utilization [59–61]. In the ﬁrst step, the CO2 decomposition was conducted using a glow discharge in the presence of hydrogen to produce CO. The catalytic hydrogenation of CO was then carried out to produce methane, liquid organic substances and soot in the second step. A similar, two-stage process has also been reported by Venugopalan and Veprek [4]. CO2 and the dissociated product CO can also be directly converted to more valuable hydrocarbons using nonthermal plasmas in the absence of a catalyst. During the 1960s, Blaustein and Fu [62] reported hydrocarbon formation from CO + H2 and CO2 + H2 using a static microwave discharge at the pressures ranging from 1.6 to 6.7 kPa. The principal products were methane and acetylene with a CO conversion up to 90%. In aiming to increase the energy yield, Mertz et al. [63] developed an improved nonstatic (ﬂow) microwave discharge reactor system, with which a CO conversion of 4–19% was achieved at pressures ranging from 1.3 to 12.0 kPa. The products were also methane and acetylene, and a reaction mechanism was presented to explain the observed discharge reactions: H2 + e → 2H + e

(10.35)

H + CO → CH + O

(10.36)

H2 + CH → CH2 + H

(10.37)

H2 + CH2 → CH4

(10.38)

CH + CH → C2H2

(10.39)

Whilst this mechanism clearly differed from the catalytic Fischer–Tropsch synthesis [27], the authors considered the generation of hydrogen radicals to be the controlling step during the formation of hydrocarbons, but a maximum CO conversion was found to occur at a 3/1 ratio of H2/CO [62, 64]. A similar but more detailed mechanism was proposed by Mach et al. [65], but the energy yield reported was too low to compete with the catalytic Fischer–Tropsch synthesis. Yoshida et al. [66] studied the inﬂuence of electrode materials on the hydrogenation of CO2, where the formed hydrocarbons included CH4, C2H4, and C2H6. At a

10.4 Hydrocarbons Synthesis from CO2

ﬁxed input dc power, the plasma current between the Fe electrodes was greater than that between the Cu electrodes. However, hydrocarbon formation was more effective on the Cu electrode, which implied that the electrode material, and in turn the electrode reactions, could also affect the plasma reactions taking place among highly excited species. Consequently, the key intermediates of COad and Had were proposed, and the special afﬁnity of CO and H to Cu was considered to be a likely reason for such high activity. 10.4.3 Higher Hydrocarbons from CH4 and CO2

Recent progress in characterizing the CO2 + CH4 plasma reactions has centered on the direct synthesis of higher hydrocarbons, using a silent discharge. In this way, besides syngas (∼40%) and light hydrocarbons (C2–C4, ∼35%), liquid fuel (C5+–C11, ∼19%), oxygenates and plasma-polymerized ﬁlm were also produced [67]. In this case, with a 3.5 s residence time operation the energy yield for light higher hydrocarbons reached 16.0 g kWh−1, for liquid fuel 5.8 g kWh−1, and for syngas 14.4 g kWh−1. A detailed analysis [68–70] of the products showed the liquid hydrocarbons to be just in the range of gasoline, with over 130 components of which 87% were highly branched and represented a high octane number. The silent discharge reaction of CH4 and CO2 represents a promising GTL route. The higher hydrocarbons produced were shown not to follow the Flory–Schulz kinetics usually found in the product distribution of the catalytic Fischer–Tropsch synthesis [27]. Rather, the chain build-up in this direct higher hydrocarbon formation using gas discharges would start from methyl radical formation. The role of CO2 in the feed would be to promote the methane conversion, although most of it would be converted into syngas and water. Two possible mechanisms have been proposed for the chain growth of higher hydrocarbons in gas discharges. The ﬁrst mechanism is that the higher hydrocarbons are produced from the chain reactions initiated by the coupling of methyl radicals generated in the discharges: CH4 + (e, H, O, O− , OH, . . . .) → CH3 + (H + e, H2, OH, OH− , H2O, . . . .)

(10.40)

The chain reactions would then proceed as follows: CH3 + CH3 → C2H6

(10.41)

C2H6 + e → C2H5 + H + e

(10.42)

C2H5 + C2H5 → C4 H10

(10.43)

C4 H10 + e → C4 H9 + H + e

(10.44)

C2H5 + C4 H9 → C6H14

(10.45)

C2H5 + CH3 → C3H8

(10.46)

The second mechanism is that the higher hydrocarbons are produced from the hydrogenation of CO, the dissociated product of CO2 [66, 71], in a way similar to

279

280

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

the plasma syngas conversion described in Equations 10.35 to 10.39. However, the ﬁrst mechanism should dominate the discharge reactions. In the presence of zeolites [68], pure methane also produced some clean liquid hydrocarbons in a DBD plasma reactor. However, to the present authors’ knowledge, there was no obvious liquid formation in other pure CH4 plasmas reaction. With CH4 and CO2 as the feed [67–70], liquid fuel was formed even in the absence of a catalyst, but the liquid was yellow due to the presence of some oxygenates. By using a heterogeneous catalyst, it is possible to control or limit the chain growth of higher hydrocarbons. The introduction of NaX and A zeolites [68, 70] into the discharge gap inhibits the formation of carbon black and plasma polymers and reduces the conversion of CH4 and CO2, but it signiﬁcantly increases the selectivities of the light and liquid hydrocarbons. A high CH4/CO2 feed ratio and a narrow discharge gap (and thus a higher mean electron energy) may favor the formation of liquid fuels. In this reaction, the liquid hydrocarbons are ignited in the gap between a highvoltage electrode covered by the dielectric material (quartz) and a grounded electrode. In another study [72], the grounded electrode was covered by the dielectric material (quartz), which meant that the high-voltage electrode would be exposed directly to plasma phase, such that the product would contain gaseous hydrocarbons, syngas, and oxygenates; however, no liquid hydrocarbons could be detected. Thus, a change in reactor conﬁguration could induce a signiﬁcant change in the chemistry involved.

10.5 Oxygenates Synthesis from CO2 10.5.1 Methanol from CO2 Hydrogenation

In a gas discharge, both CO and H can be easily excited, providing them with a sufﬁciently high energy to overcome the limitation of a high activation energy for the direct hydrogenation of CO2, even in the absence of a catalyst: CO2 + e → CO2* + e

(10.47)

H2 + e → H2* + e

(10.48)

CO2* + 3 H2 → CH3OH + H2O

(10.49)

CO2* + 4 H2* → CH4 + 2 H2O

(10.50)

or

Eliasson et al. [1, 73–77] reported a silent discharge CO2 hydrogenation in the presence or absence of a catalyst to produce methanol. A radical reaction mechanism [75] has been presented to explain the observed phenomena, and can be expressed as:

10.5 Oxygenates Synthesis from CO2

CO2 ⎯H⎯ → CO ⎯H⎯ → CHO ⎯H⎯ → CH2O ⎯H⎯ → CH3O ⎯H⎯ → CH3OH

(10.51)

The methane formation was considered as: → CH2OH ⎯H⎯ → CH3 ⎯H⎯ → CH4 CH3OH ⎯H⎯

(10.52)

The comparative experiments showed [74, 75] that the simultaneous presence of silent discharge and catalyst shifted the temperature range of the maximum catalyst activity from 493 K to 373 K. A higher pressure than atmospheric pressure is favored for methanol formation; in particular, a higher pressure will inhibit methane formation but reduce the electron energy and thereby the conversions. The presence of a catalyst was also found to increase the methanol yield and selectivity more than 10-fold compared to that obtained with discharge only, although the methanol yield obtained was still considered too low (<4%). One possible reason for this might be the existence of a reverse water gas shift reaction (Equation 10.32) that is faster than the methanol synthesis [76]. Another possible reason might be that the newly-formed methanol was not stable and thus was easily converted to other chemicals, such as dimethyl ether (DME). Quenching would be necessary to terminate any further methanol conversion in the discharges. 10.5.2 Methanol from CO2 and CH4

Based on the reactions in Equations 10.49 and 10.50, it is clear that the utilization of 1 mol of CO2 will, unfortunately, convert 1–2 mol of hydrogen into water as a waste product; consequently, the process is not attractive from an economics point of view. Consequently, attempts were made in Japan [78–80] to produce methanol from methane and CO2 directly using CO2 as an oxidant. The corona discharge and silent discharge were applied respectively, but the methanol yield remained very low, despite metal oxide catalysts being used to improve the yield. However, a signiﬁcantly higher methanol yield was achieved [81] from a 1 : 1 ratio of CO2/ CH4 mixture using a zeolite catalyst (13×) under the inﬂuence of the corona discharge; alcohol was also produced simultaneously, with concentrations of methanol and alcohol in the efﬂuents being 12.0% and 4.0%, respectively, at 100 kPa. Although a lower gas pressure reduced the methanol and alcohol concentrations in the products, no analysis of the mechanisms was reported. Thus, it was considered that the dissociation of CO2 (Equations 10.25 and 10.26) would generate certain speciﬁc oxygen species for the methanol formation from methane; among these, O(1D) was identiﬁed as one such active species [82]. According to Mallinson et al. [83], methyl radical formation also plays an important role in the methanol formation via nonthermal plasmas. Further experiments [84] conﬁrmed that methyl radicals could also induce parallel ethane formation. As CO2 and CH4 are both major greenhouse gases, it would be advantageous if it were possible to produce oxygenates (e.g., methanol) by using CO2 and CH4 as feedstocks, with a high energy yield. A discharge pathway for such CO2 and CH4

281

282

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

conversion might also be expected that would retain at least one oxygen atom of the CO2 molecule in the product. In addition to methanol, aldehydes and acids also represent this type of product. 10.5.3 Aldehydes from CH4 and CO2

The synthesis of aldehydes from CO2 and CH4 in electric discharges has been conducted by Luk’yanov et al. [85]. The effects of CO2 concentration, wall temperature, and the addition of argon, on the formation of aldehydes and carboxylic acids have also been discussed. The highest aldehyde yield achieved was 1083 μmol, at a CO2 fraction of 0.6, whilst argon behaved as a catalyst or as a dilution gas under these reaction conditions. The synthesis of aldehydes from CO2 and CH4 via a silent discharge was reported as early as the 1930s [86], in which the system consisted of a central rod electrode and a liquid-dielectric electrode. With this system, aldehydes (formaldehyde and acetaldehyde) could be readily produced even at room temperature, with the selectivity of the products being related to the CO2/ CH4 ratio. The use of equimolar proportions or an excess of methane led to a greater acetaldehyde formation, although the ﬁne detail of the conversions and yields were not reported. The formation of formaldehyde was thought to occur via methyl radicals, with the reaction described as [87]: CH3 + O → HCHO + H

(10.53)

In contrast, the newly-generated CO in Equations 10.25 and 10.26 removes extra energy from the discharge reactions, such that the CO molecules will be in an excited state and react easily with methane to generate acetaldehyde [88]: CH4 + CO → CH3CHO

(10.54)

10.5.4 Acetic Acid from CH4 and CO2

As noted above, the silent discharge reaction of CH4 and CO2 conducted by Liu and Eliasson et al. [68–71] produced liquid hydrocarbons that contained many oxygenates, the total selectivity of which was approximately 10%, with acetic acid and ethanol the two main components. It is worth noting that a signiﬁcantly high selectivity of acetic acid was obtained [68–71, 89]. Under the reaction conditions of a discharge gap of 1.1 nm, a discharge power of 100 W, a ﬂow rate of 40 ml min−1, and a CH4 fraction of 67.4%, the selectivities of acetic acid, formic acid, methanol and ethanol were 5.3%, 0.2%, 1.6%, 1.0% and 1.7%, respectively, while the conversions of methane and CO2 were 54.1% and 37.4%, respectively. Although the reaction to produce acetic acid from methane and CO2 is a perfect in terms of its atom-economy, unfortunately it is not feasible on a thermodynamic basis. CH4 + CO2 → CH3COOH, ΔG298K = 71.2 kJmol −1

(10.55)

10.5 Oxygenates Synthesis from CO2

Under the conditions of a dielectric barrier discharge, this reaction becomes true, there being two possible routes for the formation of acetic acid. The ﬁrst route is via a CO2− intermediate: CO2− + H → COOH−

(10.56)

CH3 + COOH− → CH3COOH + e −

(10.57)

while the second route is via a CO intermediate: CH3 + CO → CH3CO

(10.58)

CH3CO + O → CH3COO

(10.59)

CH3CO + O− → CH3COO− CH3COO + H → CH3COOH + e −

(10.60) −

(10.61)

or O + H → OH

(10.62)

O− + H → OH−

(10.63)

CH3CO + OH → CH3COOH

(10.64)

CH3CO + OH− → CH3COOH + e −

(10.65)

The results of density functional theory (DFT) studies [90, 91] have suggested that the energy requirement via the CO2− pathway is much less than that via the CO pathway, and thus the former is more thermodynamically favored. The methyl radical formation and dissociation of CO2 are two rate-limiting steps for the synthesis of acetic acid directly from CH4 and CO2. It is very interesting that, in the presence of a starch coating [92] on the dielectric material surface, the formation of oxygenates is signiﬁcantly enhanced. The highest selectivity was approximately 42%, with conversions of methane and CO2 of approximately 20%. The product included C1–C4 alcohols, C1–C3 acids, formaldehyde, and esters. The optimal concentration of methane for the production of oxygenates was approximately 35.4%, whilst a signiﬁcantly high selectivity of 11.2% for acetic acid was achieved with an applied power of 75 W. 10.5.5 Oxygenates from H2O and CO2

A water molecule contains two hydrogen atoms and one oxygen atom. Hydrogen and hydroxyl radicals can be produced from the dissociation [93] or dissociative electron attachment of water [94]: H2O + e → H + OH + e

(10.66)

H2O + e → OH + H

(10.67)

−

283

284

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

Hydrogen radicals generated by these reactions would induce the direct hydrogenation of CO2 and CO, the product of CO2 dissociation. Ihara et al. [95, 96] reported a microwave discharge reduction of CO2 with steam in which the principal products achieved were methanol and oxalic acid, although H2O2, was also produced in relatively large amounts. The maximum energy yields for methanol and oxalic acid were reported as 0.0333 μg W−1 and 94.2 μg W−1, respectively. A signiﬁcant improvement has been reported by Wan et al. [97, 98] on hydrocarbon formation from CO2 and H2O over a microwave-induced catalytic reaction, using supported metal catalysts. For this, when a microwave irradiation mechanism was used, the major products were CH4, CH3OH, acetone, and C3 and C4 alcohols. No hydrocarbons were detected in the absence of water, and neither were any products detected with CO2 and H2O when the irradiation was carried out in the absence of a catalyst. Based on these ﬁndings, it was suggested that an alternative and environment-friendly motor fuel might be produced from very cheap reactants of CO2 and water by using a microwave-induced catalysis, although no plasma phenomenon was mentioned in these studies.

10.6 Combination of Plasma with Catalyst 10.6.1 Catalysts in Plasma Utilization of CO2

As noted above, catalysts may be applied to many plasma reactions to improve the conversions or suppress the formation of undesired byproducts. Heterogeneous catalysts are generally packed in the plasma zone, and are in direct contact with the plasma-induced species. For example, the formation of more valuable hydrocarbons from CO2 via gas discharge in the presence of a catalyst has been reported with the multistage utilization of CO2 [4, 59–61]. In the dehydrogenation of C2H6 with CO2, rare earth oxide catalysts (e.g., CeO2 and La2O3/γ-Al2O3) were able to enhance the conversion of C2H6, whereas Pd/γ-Al2O3 favored the formation of C2H4 rather than C2H2 [52]. As a catalyst, Ni/γ-Al2O3 was reported to increase the selectivity of CO from 49.1% to 60.9% in CH4/CO2 reforming with DBD plasma [33]. Zeolites were also used in the synthesis of liquid hydrocarbons and oxygenates from CH4 and CO2 with DBD plasma; this not only allowed control of the chain propagation reaction, but also led to an inhibition of the formation of carbon black and polymers [68–70]. Another approach here would be to combine a catalytic reactor with the plasma reactor. In the case of CH4/CO2 reforming, the conversion of CH4 was 40% using DBD, and 2% using the Ni/γ-Al2O3 catalyst, with the conversion being increased to 98% when the catalytic zone was simply located behind the DBD zone [34]. In DBD reactions, it is possible to use the catalyst directly as a dielectric-barrier material. This allows an optimum process to be achieved if the relaxation of the

10.6 Combination of Plasma with Catalyst

dielectric catalyst and gas dielectric in a silent discharge reactor can be adjusted to a superior state by applying an external electric ﬁeld. Metal electrodes also play catalytic roles in some plasma reactions. For example, Cairns et al. [99] reported the formation of carbonaceous species on metallic surfaces from CO2 and methane under the inﬂuence of a RF plasma discharge. In these experiments, 20%Cr/25%Ni/Nb stabilized steel metal pins were investigated within a temperature range of 623–1023 K. The results suggested that carbonaceous deposition onto the steel surface might be induced principally by the catalytic activity of the surface itself. One ﬁnding in support of this was that the surface remained clean if the steel was coated with silica. Subsequently, Suib et al. demonstrated the activity of different metal electrodes for CO2 decomposition [19, 20]. In this case, the metals actually inﬂuenced the energy transfer from excited dilute molecules to CO2, thus adjusting the conversion of CO2. In later studies, Li et al. [36] observed similar metal-based effects in the CH4/CO2 reforming reaction using DBD, while Yoshida et al. [66] claimed that the Cu electrode was more active than the Fe electrode for the hydrogenation of CO2 to hydrocarbons. In these studies, the adsorption of CO and H onto the metals was proposed as the key step in the reaction. 10.6.2 Interaction Between Plasma and Catalyst

Recent experimental results have identiﬁed a twofold interaction between gas discharge and catalyst, namely that gas discharge promotes catalysis, while the catalyst enhances the nonequilibrium of plasmas. When Rapakoulias et al. [100] investigated heterogeneous catalysis under the inﬂuence of plasma-excited species, the surface properties of the catalyst were seen to be modiﬁed even at low gas temperatures. Such a plasma-modiﬁed catalysis would induce a signiﬁcant modiﬁcation in chemisorption and desorption, and thereby in the activity and selectivity of the catalyst. This effect of plasma on a catalyst has been used to prepare highefﬁciency catalysts for conventional catalysis. Subsequently, Kizling and Järàs [101] reviewed the use of plasma techniques in catalyst preparation, regeneration and catalytic reactions, while Liu et al. [102, 103] have summarized the more recent progress in this ﬁeld, and have regarded it as a more environment-friendly catalyst preparation method. Nonthermal plasma treatment may inﬂuence the acid–base nature of zeolites, enhance the dispersion of the supported metals, and even adjust the microstructure of the metal nanoparticles and metal–support interface. Indeed, by using nonthermal plasma, many highly active and durable catalysts have been prepared for CO2 and methane conversions. In both catalyst and plasma-combined reactions, it is expected that plasma would modify the physico-chemical properties of a catalyst in situ, after which the modiﬁed catalyst would demonstrate high performance in the plasma reaction. The presence of a catalyst in plasma regions will not only increase the reaction surface area and provide catalytic sites, but also maintain and most likely increase

285

286

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

the nonequilibrium character of gas discharge. Jogan et al. [8] have investigated the reduction of CO2 from ﬂue gas by using an ac ferroelectric packed-bed reactor. In this case, when the reactor was packed with approximately spherically shaped BaTiO3 pellets, the energy yield was higher than in the absence of BaTiO3 pellets. This effect was explained by suggesting that, if an external electric ﬁeld (usually an AC electric ﬁeld) were to be applied across the packed high-dielectric layer, the pellets would become polarized and the charge would accumulate on the dielectric surface. Consequently, an intense electric ﬁeld would be generated around each pellet contact point, resulting in microdischarges between the pellets. Such a microdischarge in the packed-bed reactor would lead to the production of energetic electrons rather than ions. Hence, gas heating would not be serious, especially in the case of a low-frequency electric ﬁeld. The packed-bed gas discharge reactor would itself behave like a silent discharge, except that the microdischarge would occur between the curved dielectric surface in the packed-bed reactor rather than on the ﬂat dielectric surface in the silent discharge. The microdischarges induce a signiﬁcant enrichment of electrons that are essential for the sustainability of plasmas. Liu et al. [104] studied the characteristics of nonthermal plasma in the presence of zeolites by using a ﬂoating, double-probe method. Here, the presence of HZSM-5 zeolite increased the electron temperature of nonthermal plasmas by 250%, while the discharge power was reduced by 58%. This might lead to a higher reactivity and energy efﬁciency of plasma catalytic reactions. It is known that the framework of catalysts such as zeolites can create intense natural electric ﬁelds (up to 1 V Å−1), which is much stronger than any electric ﬁeld applied for the initiation of man-made plasmas. Even without a man-made plasma, some catalysts would already be in a “natural gas discharge” status. The only difference between the “natural gas discharge” of a catalyst and the manmade plasmas, is the size of the “gas discharge” region. Man-made plasmas are in the macroscale, and normally generated by dc or ac high voltages; in contrast, the “gas discharge” created by a catalyst defect or by a zeolite framework is in the nanoscale, with a high-ﬁeld character. Typically, almost all investigations in plasma chemistry have employed high voltages to generate plasmas, irrespective of the type of plasma applied. Many excited species, with relatively lower energies, do not contribute to the syntheses but rather consume much energy and decrease selectivity. The analogy between gas discharge plasmas and the environment within the catalyst pores suggests that a shift of plasma chemistry, from highvoltage plasmas to high-ﬁeld plasmas, may provide signiﬁcantly higher efﬁciencies [103].

10.7 Summary

The potential of nonthermal plasma technologies for the utilization of CO2 has been addressed in this chapter. Nonthermal plasma pathways have many advantages over conventional catalysis, notably that they can be generated under ambient

References

conditions and require no complex systems. Nonthermal plasma covers almost all mechanisms of modern synthetic chemistry, including electrochemistry, photochemistry, radiation chemistry, and homogeneous and heterogeneous catalysis. Based on the characteristics of nonthermal plasmas, a selective product would be very expensive to create; hence, an ideal but feasible product via nonthermal plasmas may be a mixture of higher hydrocarbons and oxygenates, with syngas as a byproduct when using methane as a cofeed. In principle, whilst plasma CO2 conversion is highly effective and efﬁcient, it has a low energy efﬁciency and can only be applied when cheap or CO2-free electricity is available. Thus, if the plasma utilization of CO2 is to become economically attractive, further investigations – not only into the mechanisms of nonthermal plasma or catalytic plasma but also on plasma generator and process technology – are absolutely necessary.

Acknowledgments

The authors gratefully acknowledge support from ABB Switzerland Ltd., the Ministry of Science and Technology of China (under contracts 2005CB221406 and G1999022402), the National Natural Science Foundation of China (under contract 20490203), and the Ministry of Education of China (under contract 0212).

References 1 Eliasson, B., Simon, F.G., and Egli, W. (1993) Non-thermal Plasma Techniques for Pollution Control, vol. G34 (eds B.M. Penetrante and S.E. Schultheis), NATO ASI Series, Part B, Springer-Verlag, pp. 321–337. 2 Aresta, M. and Dibenedetto, A. (2004) Catal. Today, 98, 455–462. 3 Liu, C.-J., Xu, G., and Wang, T. (1999) Fuel Process. Technol., 58, 119–134. 4 Venugopalan, M. and Veprek, S. (1983) Topics in Current Chemistry: Plasma Chemistry IV (ed F.L. Boschke), Springer-Verlag, New York. 5 Eliasson, B. and Kogelschatz, U. (1991) IEEE Trans. Plasma Sci., 19, 1063–1077. 6 Grill, A. (1994) Cold Plasma in Materials Fabrication: From Fundamentals to Applications, IEEE Press, New York. 7 Biederman, H. and Osada, Y. (1991) Plasma Polymerization Processes, Elsevier. 8 Maezono, I. and Chang, J.-S. (1990) IEEE Trans. Ind. Appl., 29, 651–655. 9 Czernichowski, A. (1994) Pure Appl. Chem., 66, 1301–1310.

10 Kennedy, L.A., Fridman, A., and Saveliev, A. (1996) 49th Annual Meeting of Division of Fluid Dynamics of the American Physical Society, November 24–26, Syracuse, New York. 11 Fridman, A. (1993) Proceedings of the 11th International Symposium on Plasma Chemistry, August 22–27, Loughborough, England, pp. 257–262. 12 Pellerin, S., Cormier, J.-M., Chapelle, J., and Kassabji, F. (1997) Proceedings of the 13th International Symposium on Plasma Chemistry, August 18–22, Beijing, pp. 813–818. 13 Lin, L. and Wu, C. (1997) Proceedings of the 13th International Symposium on Plasma Chemistry, August 18–22, Beijing, pp. 807–812. 14 Chang, J.-S. (1987) Proceedings of International Symposium on High Pressure Low Temperature Plasma Processing, IEE Japan Press, pp. 45– 54. 15 Carabine, M.D., and Vasanthakumar, L.S. (1995) Proceedings of the 12th International Symposium on Plasma

287

288

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches

16

17

18

19

20

21

22

23

24

25

26

27 28

29

30

Chemistry, August 21–25, Minneapolis, pp. 1071–1076. Boukhalfa, N., Goldman, A., Goldman, M., and Sigmond, R.S. (1972) J. Chem. Phys., 57, 3180–3186. Beuthe, T.G. and Chang, J.-S. (1995) Proceedings of the 12th International Symposium on Plasma Chemistry, August 21–25, Minneapolis, pp. 717–772. Liu, C.-J., Mallinson, R., and Lobban, L. (1998) Proceeding of the 4th International Conference on Greenhouse Gas Control Technologies (GHGT-4), August 30–September 2. Brock, S.L., Marquez, M., Suib, S.L., Hayashi, Y., and Matsumoto, H. (1998) J. Catal., 180, 225–233. Suib, S.L., Brock, S.L., Marquez, M., Luo, J., Matsumoto, H., and Hayashi, Y. (1998) J. Phys. Chem. B, 102, 9661–9666. Wang, J.-Y., Xia, G.-G., Huang, A., Suib, S.L., Hayashi, Y., and Matsumoto, H. (1999) J. Catal., 185, 152–159. Andreev, S.N., Zakharov, V.V., Ochkin, V.N., and Savinov, S.Y. (2004) Spectrochim. Acta Part A, 60, 3361– 3369. Li, R., Tang, Q., Yin, S., Yamaguchi, Y., and Sato, T. (2004) Chem. Lett., 33, 412–413. Li, R., Tang, Q., Yin, S., and Sato, T. (2006) Fuel Process. Techn., 87, 617– 622. Li, R., Tang, Q., Yin, S., and Sato, T. (2006) Plasma Chem. Plasma Process., 26, 267–276. Indarto, A., Yang, D.R., Choi, J.-W., Lee, H., and Song, H.K. (2007) J. Hazard. Mater., 146, 309–315. Biloen, P. and Sachtler, W.M.H. (1981) Adv. Catal., 30, 165–216. Gesser, H.D., Hunter, N.R., and Probawono, D. (1998) Plasma Chem. Plasma Process., 18, 241–245. Kogelschatz, U., Zhou, L.-M., Xue, B., and Eliasson, B. (1998) Proceeding of the 4th International Conference on Greenhouse Gas Control Technologies (GHGT-4), August 30–September 2. Kogelschatz, U., and Kraus, M. (1998) Proceeding of the 4th International

31

32 33

34 35

36

37

38

39

40 41

42

43

44

45

46

47

Conference on Greenhouse Gas Control Technologies (GHGT-4), August 30–September 2. Motret, O., Pellerin, S., Nikravech, M., Massereau, V., and Pouvesle, J.M. (1997) Plasma Chem. Plasma Process., 17, 393–407. Tsang, S.C., Claridge, J.B., and Green, M.L.H. (1995) Catal. Today, 23, 3–15. Song, H.K., Choi, J.-W., Yue, S.H., Lee, H., and Na, B.-K. (2004) Catal. Today, 89, 27–33. Lee, H., Song, H.K., and Min, B.R. (2006) Chem. Lett., 35, 646–647. Lee, H., Lee, C.-H., Choi, J.-W., and Song, H.K. (2007) Energy Fuels, 21, 23–29. Li, Y., Xu, G., Liu, C., Eliasson, B., and Xue, B. (2001) Energy Fuels, 15, 299–302. Sarmiento, B., Brey, J.J., Viera, I.G., González-Elipe, A.R., Cotrino, J., and Rico, V.J. (2007) J. Power Sources, 169, 140–143. Li, M., Xu, G., Tian, Y., Chen, L., and Fu, H. (2004) J. Phys. Chem. A, 108, 1687–1693. Li, M.-W., Liu, C.-P., Tian, Y.-L., Xu, G.-H, Zhang, F.-C., and Wang, Y.-Q. (2006) Energy Fuels, 20, 1033–1038. Li, M.-W., Tian, Y.-L., and Xu, G.-H. (2007) Energy Fuels, 21, 2335–2339. Yao, S.L., Ouyang, F., Nakayama, A., Suzuki, E., Okumoto, M., and Mizuno, A. (2000) Energy Fuels, 14, 910–914. Yao, S.L., Okumoto, M., Nakayama, A., and Suzuki, E. (2001) Energy Fuels, 15, 1295–1299. Huang, A., Xia, G., Wang, J., Suib, S.L., Hayashi, Y., and Matsumoto, H. (2000) J. Catal., 189, 349–359. Zhang, J.-Q., Zhang, J.-S., Yang, Y.-J., and Liu, Q. (2003) Energy Fuels, 17, 54–59. Lesueur, H., Czernichowski, A., and Chapelle, J. (1994) Int. J. Hydrogen Energy, 19, 139–144. Czernichowski, A. (1998) Proceeding of the 4th International Conference on Greenhouse Gas Control Technologies (GHGT-4), August 30–September 2. Indarto, A., Choi, J.-W., Lee, H., and Song, H.K. (2006) Energy, 31, 2986– 2995.

References 48 Ghorbanzadeh, A.M., Norouzi, S., and Mohammadi, T. (2005) J. Phys. D: Appl. Phys., 38, 3804–3811. 49 Futamura, S. and Annadurai, G. (2005) IEEE Trans. Ind. Appl., 41, 1515–1521. 50 Sugasawa, M. and Futamura, S. (2008) IEEE Trans. Ind. Appl., 44, 46–52. 51 Futamura, S., and Annadurai, G. (2008) IEEE Trans. Ind. Appl., 44, 53–60. 52 Zhang, X., Zhu, A., Li, X., and Gong, W. (2004) Catal. Today, 89, 97–102. 53 Futamura, S. and Kabashima, H. (2004) Carbon Dioxide Utilization for Global Sustainability: Proceedings of the 7th International Conference on Carbon Dioxide Utilization, Seoul, Korea, October 12–16, 2003, pp. 119–124. 54 Potapkin, B.V., Strelkova, M.I., Fridman, A.A., Harkness, J.B.L., and Doctor, R.D. (1995) Proceedings of the 12th International Symposium on Plasma Chemistry, August 21–25, Minneapolis, pp. 1737–1742. 55 Nakagawa, Y., Kohchi, A., and Nishitani, T. (1993) Jpn. J. Appl. Phys., 32, L1568–L1570. 56 Krylov, O.V. and Kh Mamedov, A. (1995) Heterogeneous catalytic reactions of carbon dioxide. Russ. Chem. Rev., 64, 877–900. 57 Fotouh, K., and Liu, C.-J. (1996) Preprints of the 212th ACS National Meeting, August 26, vol. 41, pp. 1382–1386. 58 Liu, C.-J. (1996) Chem. Lett., 749–750. 59 Eremin, E.N., Mal’tsev, A.N., Belova, V.M., and Ivanter, V.L. (1978) Russian J. Phys. Chem., 52, 1654–1656. 60 Eremin, E.N., Mal’tsev, A.N., Belova, V.M., and Ivanter, V.L. (1979) Russian J. Phys. Chem., 53, 1159–1161. 61 Eremin, E.N., Mal’tsev, A.N., Belova, V.M., and Ivanter, V.L. (1979) Russian J. Phys. Chem., 53, 1448–1450. 62 Blaustein, B.D. and Fu, Y.C. (1969) Le Chatelier’s Principle in Discharge Reactions, vol. 80 (ed. R.F. Gould), Advances in Chemistry Series, American Chemical Society, Washington, DC. 63 Mertz, S.F., Asmussen, J., and Hawley, M.C. (1974) IEEE Trans. Plasma Sci., 2, 297–306.

64 Lett, R.G., Steiner, W.A., and Fu, Y.C. (1972) J. Phys. Chem., 76, 2941–2946. 65 Mach, R., Drost, H., Dube, G., and Spangenberg, H.-J. (1983) Beitr. Plasmaphys., 23, 595–602. 66 Yoshida, Z., Yosue, H., and Nogami, G. (2001) J. Electrochem. Soc., 148, D55– D59. 67 Eliasson, B., Liu, C.-J., and Killer, E. (2001) US Patent 6326407 B1. 68 Eliasson, B., Liu, C.-J., and Kogelschatz, U. (2000) Ind. Eng. Chem. Res., 39, 1221–1227. 69 Liu, C.-J., Xue, B., Eliasson, B., He, F., Li, Y., and Xu, G.-H. (2001) Plasma Chem. Plasma Process., 21, 301–310. 70 Jiang, T., Li, Y., Liu, C., Xu, G., Eliasson, B., and Xue, B. (2002) Catal. Today, 72, 229–235. 71 Chantry, P.J. (1972) J. Chem. Phys., 57, 3180–3186. 72 Zhang, Y., Li, Y., Wang, Y., Liu, C., and Eliasson, B. (2003) Fuel Process. Techn., 83, 101–109. 73 Eliasson, B., Kogelschatz, U., Killer, E., and Bill, A. (1996) Proceedings of the 11th World Hydrogen Energy Conf. (Hydrogen ‘96), Stuttgart, Germany, June 23–28, pp. 2449–2459. 74 Bill, A., Eliasson, B., Kogelschatz, U., and Zhou, L.-M. (1998) Stud. Surf. Sci. Catal., 114, 541–544. 75 Eliasson, B., Kogelschatz, U., Xue, B.Z., and Zhou, L.M. (1998) Ind. Eng. Chem. Res., 37, 3350–3357. 76 Bill, A., Wokaun, A., Eliasson, B., Killer, E., and Kogelschatz, U. (1997) Energy Convers. Mgmt, 38 (Suppl.), S415–S422. 77 Kogelschatz, U., and Eliasson, B. (1996) Presentation on the International Symposium on high pressure low temperature plasma chemistry, Milovy, Czech Republic, September 2–4. 78 Rajanikanth, B., Okumoto, M., Katsura, S., and Mizuno, A. (1996) Proceedings of 1996 IEEE IAS Annual Meeting, pp. 1813–1817. 79 Okazaki, K. (1993) Proceedings of the International Symposium on CO2 Fixation and Efﬁcient Utilization of Energy, Tokyo, Japan, pp. 37–42. 80 Indarto, A. (2008) IEEE Trans. Dielectric Electr. Insulation, 15, 1038–1043.

289

290

10 Utilization of Carbon Dioxide through Nonthermal Plasma Approaches 81 Helmut, D., Reinhard, M., and Jürgen, R. (1988) Deutsche Demokratische Republik Patent DD 260 011 A1. 82 Parnis, J.M., Hoover, L.E., Pederson, D.B., and Patterson, D.D. (1995) J. Phys. Chem., 99, 13528–13536. 83 Mallinson, R.G., Sliepcevich, C.M., and Rusek, S. (1987) Preprints Am. Chem. Soc. Fuel Chem. Div., 32, 266. 84 Bhatnagar, R. and Mallinson, R.G. (1995) Methane and Alkane Conversion Chemistry (eds M.M. Bhasin and D.W. Slocue), Plenum, New York, pp. 249–264. 85 Luk’yanov, V.B. and Eremeev, A.P. (1974) Russian J. Phys. Chem., 48, 531–533. 86 Finlayson, D. and Geoffrey, J. (1935) US Patent 1986885. 87 Oumghar, A., Legrand, J.C., Diamy, A.M., and Turillon, N. (1995) Plasma Chem. Plasma Process., 15, 87– 107. 88 Glocker, G. and Lind, S.C. (1939) The Electrochemistry of Gases and Other Dielectrics, John Wiley & Sons, Inc., New York, p. 193. 89 Liu, C., Li, Y., Zhang, Y., Wang, Y., Zou, J., Eliasson, B., and Xue, B. (2001) Chem. Lett., 1304–1305. 90 Wang, J., Liu, C., and Eliassion, B. (2004) Energy Fuels, 18, 148– 153. 91 Wang, J., Liu, C., Zhang, Y., and Eliasson, B. (2003) Chem. Phys. Lett., 368, 313–318. 92 Zou, J.-J., Zhang, Y., Liu, C.-J., Li, Y., and Eliasson, B. (2003) Plasma Chem. Plasma Process., 23, 69–82.

93 Garcia, M. and Chang, B. (1995) Proceedings of the 12th International Symposium on Plasma Chemistry, August 21–25, Minneapolis, pp. 813–818. 94 Penetrante, B.M. (1993) Non-thermal Plasma Techniques for Pollution Control, vol. G34 (eds B.M. Penetrante and S.E. Schultheis), NATO ASI Series, Part A, Springer-Verlag. 95 Ihara, T., Kibobu, M., and Iriyama, Y. (1994) Bull. Chem. Soc. Jpn, 67, 312–314. 96 Ihara, T., Ouro, T., Ochiai, T., Kibobu, M., and Iriyama, Y. (1996) Bull. Chem. Soc. Jpn, 69, 241–244. 97 Wan, J.K.S., Bamwenda, G., and Depew, M.C. (1991) Res. Chem. Intermediates, 16, 241–255. 98 Wan, J.K.S. (1993) Res. Chem. Intermediates, 19, 147–158. 99 Cairns, J.A., Coad, J.P., Richards, E.W.T., and Stenhouse, L.A. (1980) Nature, 288, 686–687. 100 Rapakoulias, D.E., Cavadias, S., and Mataras, D. (1993) High Temp. Chem. Processes, 2, 231–246. 101 Kizling, M.B. and Järàs, S.G. (1996) Appl. Catal. A, 147, 1–21. 102 Liu, C., Vissokov, G.P., and Jang, B.W.-L. (2002) Catal. Today, 72, 173–184. 103 Liu, C., Zou, J., Yu, K., Cheng, D., Han, Y., Zhan, J., Ratanatawanate, C., and Jang, B.W.-L. (2006) Pure Appl. Chem., 78, 1227–1238. 104 Liu, C., Wang, J., Yu, K., Eliasson, B., Xia, Q., Xue, B., and Zhang, Y. (2002) J. Electrostatics., 54, 149–158.

291

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction of Carbon Dioxide Emily Barton Cole and Andrew B. Bocarsly

11.1 Introduction

The reduction of carbon dioxide (CO2) has received a good deal of attention in recent years. With increasing concerns about rising atmospheric CO2 levels, scientists have discussed new strategies to reduce the impact of CO2 on global warming. Many ideas have involved trapping the “greenhouse gas” and converting it into fuels and organic materials, using either light or electrical energy [1–3]. The idea of synthesizing valuable organics from CO2, however, is not new. Some of the earliest reports of CO2 reduction date back to the late 1800s, when formic acid was ﬁrst synthesized from aqueous bicarbonate [4]. Since then, there has been increasing interest in the ﬁelds of the photochemical, electrochemical, and photoelectrochemical reduction of CO2 to create useful products. In this chapter, the major early studies, as well as more recent advances in these ﬁelds, will be examined. As the reduction of CO2 is energetically uphill, this process involves a transfer of energy, usually derived either from light or electricity. The earliest studies in this ﬁeld examined metal electrodes, in electrochemical systems, which were capable of reducing CO2 through an applied bias or current. Achievements in the ﬁelds of photochemistry and photoelectrochemistry during the 1970s sparked an increased interest in the possibility of using sunlight to reduce CO2, when research groups began to study CO2 reduction by using a variety of semiconductor electrodes and photosensitizers, initially derived from efforts to reduce H2O to H2. These achievements stemmed primarily from the search for alternative energy sources, as energy costs dramatically increased in the western world due to the Arab oil embargo during the early 1970s. The number of publications related to CO2 reduction peaked in the early 1990s, and have decreased since. Recently, rising CO2 levels, which are thought to play a role in global warming, along with increasing fuel costs, have directed more attention to the reduction of CO2. The seemingly established ﬁelds of photochemistry, electrochemistry, and photoelectrochemistry are, in effect, experiencing a revival in relation to energy science. The aim here is to present some of the major achievements that have been made in these ﬁelds, while demonstrating that ample investigations have still Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

292

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

to be conducted to develop actual industrial-scale schemes for CO2 mitigation and utilization. In this regard, the ﬁelds are all very young, with major breakthroughs on the horizon.

11.2 Homogeneous Photochemical Reduction

Homogeneous photochemical systems involve solubilized molecules that are able to convert light energy into stored chemical energy in the form of reduced CO2. The molecular photosensitizer ﬁrst absorbs photons, creating a reactive excited state that is capable of reducing CO2 through a transfer of charge. After this charge transfer, the oxidized photosensitizer receives an electron to regenerate the initial ground-state species and complete the catalytic cycle. There are two different schemes involving molecular photosensitizers. In one scheme, the photosensitizer simply serves to capture the incident light energy and then transfers this energy to a separate cocatalyst, or electron mediator, which actually completes the CO2 reduction step. In the second scheme, the photosensitizer also serves as the catalyst, directly reducing CO2 in a single step. In either route, the oxidized photosensitizer must receive an electron to regenerate the initial molecule. The addition of various quenching agents, or sacriﬁcial reductants, reduces the oxidized photosensitizer back to its ground state via two different routes. In the ﬁrst route the quenching agent simply supplies an electron to the oxidized species at the end of the catalytic cycle. This can be referred to as direct quenching, and is usually what occurs in systems where the excited state reduction potential is sufﬁcient to reduce CO2. In systems where the excited state reduction potential is not sufﬁcient to reduce CO2, a highly reducing quenching agent can be used initially to generate a higher-energy ground-state photosensitizer molecule. This is often referred to as reductive quenching, as the quenching agent generates a more reducing form of the photosensitizer which then initiates the catalytic cycle. In both schemes, the quenching agent is oxidized and allows the catalytic cycle to be completed. In this section, the various transition metal complexes, macrocycles, and aromatic hydrocarbons which act as either photosensitizers or cocatalysts for the reduction of CO2, will be described. As noted above, the early work in the ﬁeld, as well as the more recent achievements of the past 10 years, will also be examined. For further information on transition metal and macrocycle complexes which catalyze the reduction of CO2, the reader is directed to various reviews [5–8]. 11.2.1 General Considerations

In both aqueous and nonaqueous photosensitizer systems, typically the main products are the two-electron reduced CO2 species, namely CO and formate. Both

11.2 Homogeneous Photochemical Reduction

products seem to be formed through various metal–CO2 complex intermediates, derived from either the photosensitizer or the cocatalyst. As a ligand, there is crystallographic evidence showing CO2 bonded: (i) through carbon, η1; and (ii) roughly between the carbon atom and one of the oxygen atoms, η2 [9, 10]. Bonding through oxygen alone, however, has not been observed. Extensive reviews of CO2 bonding to transition metal centers, and the mechanisms of CO2 reduction, have been prepared by Creutz, Sutin et al., and by Keene et al. [9–11]. At this point, the proposed pathways to CO and formate through metal–CO2 complex intermediates will be brieﬂy discussed. Two different metal–CO2 complex intermediates have been proposed for the production of CO–metallocarboxylates and metal formates. The difference between the two species is based on the site of protonation, at the carbon atom in metallocarboxylates and at one of the oxygen atoms in metal formates. Carbonprotonation has not been observed experimentally, while oxygen-protonation is well known [9]. Isomerization can occur between metallocarboxylates and metal formates, and loss of a hydroxide group from the metal formate species yields the M–CO complex. Similarly, the direct reaction of metal complexes with free, dissolved CO2 has also been described. In this mechanism, the metal complex reacts with an oxide acceptor, such as CO2, generating the metal–CO complex and CO32− [9]. Metal formates can also result from the direct insertion of CO2 into a metal hydride bond. Speciﬁcally, in cobalt and nickel macrocycles (where the products are usually CO and H2) there is evidence of an intermediate hydride species. The mechanisms of CO2 activation by these complexes have been reviewed previously [12, 13]. One example of CO2 reduction to CO through metal hydride insertion was proposed by Lehn and Ziessel for the transition metal photocatalyst complex [Ru(bpy)2(CO)H]+ [14]. Here, the complex is ﬁrst reduced by one electron to the reactive intermediate [Ru(bpy)2(CO)H], which is activated toward CO2 insertion into the metal hydride bond; this yields the metal formate species. This complex is then reduced by a second electron, displacing formate with an acetonitrile solvent molecule to give [Ru(bpy)2(CO)CH3CN]. The reaction of this intermediate with CO2 and H2O then regenerates the starting catalyst, displacing the acetonitrile molecule and yielding HCO3− . Formate production stems from similar metal–CO2 intermediate species that yield CO as a product. Formate can be formed by the protonation of metal–CO2 complexes through intermediates that have not been determined experimentally, namely the metallocarboxylate intermediate described above. A proposed mechanism for formate production by transition metal complexes also involves a metal hydride intermediate, where CO2 actually inserts into the metal hydride bond to form the metallocarboxylate intermediate [9]. The reaction of the CO2 reduction products, however, with the metal centers for both transition metal and macrocyclic complexes, commonly leads to their deactivation. Speciﬁcally, the coordination of CO to the metal centers forms carbonato complexes, deactivating the complex to further CO2 reduction. The deactivation pathways typically result in the formation of carbonato or bicarbonate

293

294

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

complexes. Thus, there is interest in catalysts that have little afﬁnity for CO ligand coordination such as the photocatalyst Re(CO)3(bpy)Cl. However, this photocatalyst has been observed to become deactivated when it forms a formate complex (see below) [15]. Most photosensitizers, however, are reasonably photostable compounds, and their optical properties have been studied in depth. In particular, there has been much interest in ruthenium-based photosensitizers such as [Ru(bpy)3]2+ and [Ru(phen)3]2+, due to their stability and absorption of visible light. Detailed information on their optical properties, including ground and excited state information in relation to photosensitization, has been reviewed by Creutz et al. [16]. Similarly, the photochemistry and photophysics of rhenium complexes, as discussed here, have been reviewed in detail by Kirgan et al. [7]. 11.2.2 Transition Metal Complexes 11.2.2.1 Ruthenium Complexes The photosensitizer [Ru(bpy)3]2+ was examined for the reduction of CO2 and water in early studies by Lehn and Ziessel [17]. The complex [Ru(bpy)3]2+ does not interact directly with CO2, rather it serves as a chromophore and transfers energy to a cocatalyst. Under visible light illumination, together with CoCl2 as the cocatalyst, CO and H2 were produced in a CO2-saturated solution of water/ acetonitrile/triethylamine (TEA), where TEA served as the sacriﬁcial reductant. Interestingly, cobalt’s activity was unique compared to other metals. Also, the selectivity for CO could be enhanced by replacement of TEA with triethanolamine (TEOA). The mechanism was proposed to pass through a reductive quenching, with the enhanced ground-state complex reducing Co(II) to Co(I), which could then reduce CO2. Further experiments showed that the active Co complex was actually [Co(bpy)n]+, formed by a loss of ligands from the ruthenium complex [18]. The importance of this intermediate in interacting with CO2 was shown in an additional experiment, when the addition of phosphines and dimethylglyoxime as ligands stopped CO production, suggesting the requirement for an open site on the Co(I) species to which CO2 can bind [19]. Ishida et al. also used [Ru(bpy)3]2+ as the photosensitizer and [Ru(bpy)2(CO)2]2+ as the cocatalyst to produce CO and formate in a water/dimethylformamide (DMF) mixture with 1-benzyl-1,4-dihydronicotinamide (BNAA) as the sacriﬁcial reductant [20]. Changing the quenching agent to TEOA gave a higher solution pH and yielded primarily formate, with a maximal quantum yield of 14%. At the lower pH of the BNAA solutions, the metal–CO2 complex was easily protonated, forming CO and water. In aqueous solution, [Ru(bpy)3]2+ was used as the photosensitizer and methyl viologen as the electron mediator, with TEOA or EDTA as the sacriﬁcial reductant [21]. Formic acid was produced with a quantum yield of ∼1%. However, later studies by Hawecker et al. showed that methyl viologen was not the active catalyst, but rather a species formed from the decomposition of [Ru(bpy)3]2+, the [Ru(bpy)2(CO)H]+ intermediate described in Section 11.2.1 [22]. Lehn and Ziessel

11.2 Homogeneous Photochemical Reduction

similarly observed upwards of 15% quantum efﬁciency for formate with their best mixture of [Ru(bpy)3]2+ and [Ru(bpy)2(CO)H]+ [14]. 11.2.2.2 Rhenium Complexes Early studies on rhenium complexes of the form Re(CO)3(bpy)X, where X = Cl, Br, were observed to reduce CO2 to CO. The complexes acted as both the photosensitizer and catalyst, reducing CO2 in one step. Kutal et al. observed a quantum yield of 15% for CO production using the complex Re(CO)3(bpy)Br in a CO2saturated DMF/TEOA solution under 436 nm irradiation [23]. In a similar system, the Re(CO)3(bpy)Cl photosensitizer gave a quantum yield of 14% for CO [15]. However, quenching studies showed that CO2 did not interact directly with the excited state complex, and CO2 showed no increased quenching in comparison to Ar [23]. The reduced intermediate is thought to lose CO·, allowing either CO2 or H+ to bind together with an electron. The Re–H complex could then yield H2 (small amounts were observed in the absence of CO2), or could insert CO2 to give an oxygen-bound formate complex, as described above. The binding of CO2 would yield CO and water by protonation of the intermediate, thus reforming the Re(CO)3(bpy)X complex [15]. Formation of the inactive formate complex seemed to cause deactivation of the photocatalytic system [24, 25]. To increase the amount of light absorbed, [Ru(bpy)3]2+ has been added as a photosensitizer in the Re(CO)3(bpy)Cl system; however, this system was less selective for CO production, also generating H2 [26]. Hori et al., when using the photocatalyst {Re(bpy)(CO)3[P(OEt)3]}SbF6, achieved a quantum yield of 38% for CO in a CO2-saturated DMF/TEOA solution illuminated with 365 nm light [27]. Under the same conditions, the quantum yield for CO using Re(CO)3(bpy)Cl was only 16%, but comparable to that in previous reports. In more recent years, quantum efﬁciencies of 16–20% have been reported for CO production using Re ( 4, 4 ′-X 2 bpy )(CO)3PR 3+ photocatalysts, where X = H, CH3; R = P(OC2H5)3, P(O-i-C3H7)3 [28]. The solution was DMF/TEOA, irradiated at 365 nm, and the one-electron reduced species was observed to interact with CO2. Similarly, Hayashi et al. examined tricarbonyl Re(I) complexes for the reduction of CO2 to CO in dry DMF under illumination with a 150 W xenon lamp ﬁltered to emit wavelengths >380 nm [29]. For their [Re(dmb)(CO)3]2 complex (where dmb = 4,4′-dimethyl-2,2′-bipyridine), it was possible to produce CO with 25–50% yield based on the Re complex concentration. An observed CO2-bridged dimer was thought to be an intermediate in CO production; however, dimerization of the Re complex alone also competed with CO2 reduction. Both, [Re(dmb)(CO)3]2(OCO2) and Re(dmb)(CO)3(OC(O)OH), were detected as oxidation products; however, a similar formato-rhenium species to that discussed above was not detected. Hori et al. studied the photocatalysts Re(bpy)(CO)3Cl and [Re(bpy)(CO)3(POiPr)3]+ in a high-pressure DMF/amine system for the production of CO under 356 nm illumination [30, 31]. The high-pressure system (2.45 Pa CO2) resulted in a 5.1-fold higher turnover number (TON) than was achieved under normal pressure (0.1 Pa) for the Re(bpy)(CO)3Cl catalyst. The [Re(bpy)(CO)3(POiPr)3]+ photocatalyst showed a 3.8-fold increase at 3.8 Pa. The TONs for CO production peaked at 42 and 16 for each catalyst, respectively.

295

296

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

Recently, Ru and Re bi-nuclear (Ru–Re) and tetranuclear (Ru–Re3) complexes for the photoreduction of CO2 have been synthesized [32]. Under irradiation at 480 nm, in a DMF/TEOA solution, the complexes were observed to undergo reductive quenching by 1-benzyl-1,4-dihydronicotinamide (BNAH). The best bi-nuclear complex yielded a quantum efﬁciency of 9% for CO production, and a TON of 170. Similarly, the tetranuclear complex yielded a CO quantum efﬁciency of 12% and a TON of 240. While these marked increases in TONs are encouraging, it should be stated that increases of many more orders of magnitude are required to yield an economically viable system. 11.2.3 Macrocyclic Complexes

Much interest has also been expressed in tetra-azamacrocyclic compounds, due to their role in the natural reduction of CO2 to CH4 by a nickel tetrapyrrole coenzyme found in methane-producing bacteria. Tinnemans et al. used Co(II) tetra-azamacrocyclic complexes with [Ru(bpy)3]2+ as the photosensitizer and ascorbic acid as the sacriﬁcial electron donor in aqueous CO2-saturated solutions at acidic pH [33]. Whilst the TON for the total observed products of CO and H2 exceeded 500, they were formed in a ratio of 0.27 : 1, respectively. Grant et al. studied a similar system using [Ni(cyclam)]2+ as the catalyst (cyclam = 1,4,8,11-tetra-azacyclo tetradecane), [Ru(bpy)3]2+ as the photosensitizer, and ascorbic acid as the sacriﬁcial reductant [34], and observed a pH dependence on CO/H2 ratios, with the best ratio of 0.83 : 1 at pH 5. When Kimura et al. prepared pyridine derivatives of [Ni(cyclam)]2+ [35], the best complex, in CO2-saturated ascorbate buffer at pH 5.1 and [Ru(bpy)3]2+ as the photosensitizer, produced 5.8-fold more CO than [Ni(cyclam)]2+. Mochizuki et al. synthesized a bimacrocyclic Ni(II) complex, [6,6′-bi(5,7dimethyl-1,4,8,11-tetraazacyclotetradecane)]-dinickel(II) triﬂate [36]. In a CO2saturated aqueous solution at pH 4, with [Ru(bpy)3]2+ as the photosensitizer, and ascorbic acid as the sacriﬁcial reductant, the CO/H2 ratio was 15 : 1 and the rate of CO production was approximately eightfold higher than for [Ni(cyclam)]2+. Matsuoka used a different photosensitizer, p-terphenyl, with a cobalt(III) cyclam as the catalyst [37–39]. In a CO2-saturated acetonitrile/methanol solution with either TEOA or TEA as the sacriﬁcial reductant, the quantum efﬁciencies for CO and formic acid production were 15% and 10%, respectively, under 313 nm illumination. Again, however, the TONs and production rates for macrocyclic complexes were low.

11.3 Electrochemical Reduction

Electrochemical systems for the reduction of CO2 require the application of an external bias or current to supply the electrons to reduce CO2. Instead of a sacri-

11.3 Electrochemical Reduction

ﬁcial reductant, the corresponding oxidation occurs at a counterelectrode some distance from the cathode. However, the cathodic and anodic reactions must be separated using a two-compartment cell so that the reduced CO2 species are not reoxidized. Although a wide variety of metal electrodes have been examined for the reduction of CO2, only some of the more major advances in CO2 reduction at metal electrodes are highlighted here. More extensive surveys of the electrochemical reduction of CO2 are provided elsewhere [40–44]. First, CO2 reduction at metal electrodes in both aqueous and nonaqueous media, as well as in systems coupled with electron-mediating complexes are detailed. The faradaic efﬁciency of such a system can be used as a measure of efﬁciency and selectivity. For a speciﬁc, electrochemically generated product, the faradaic efﬁciency is the ratio of the actual and theoretical amounts of product formed within the same time interval, based on charge passed. An efﬁcient and selective system will lead to a 100% faradaic yield for a single product; in other words, all of the charge passed in the system has gone into the production of that product. The photochemical systems discussed in Section 11.2 yield chemistry that is limited to the two-electron reduction of CO2 to form CO or formate. Although electrochemical reduction provides a route to more highly reduced species, the kinetics of both aqueous and nonaqueous electrochemical systems are very sluggish. Speciﬁcally, the ﬁrst electron reduction of CO2 to the radical anion, iCO2− , is very energy-intensive, in part because the added electron must occupy the lowest energy π* orbital of CO2, forcing a geometric change in the molecule. The subsequent reduction steps of CO2 to highly reduced species require multiple electron transfers that are usually coupled to protonation, or proton-coupled electron transfer. The net effect is that the observed currents are usually small, and the overpotentials from the thermodynamic CO2 reduction potentials abnormally large – two limiting factors when constructing commercially viable systems. An aqueous solution is advantageous in that it provides a readily available proton source, but competition of CO2 reduction with hydrogen production then becomes a major issue. Nonaqueous solutions offer the advantage of increased CO2 solubility over aqueous media, and yield somewhat higher current densities; however, usually the less economically useful CO is produced. Electrochemical systems that produce high faradaic efﬁciencies for highly reduced products and good current densities at modest electrode potentials remain a challenge. 11.3.1 Reduction in Aqueous Solutions at Metal Electrodes

In aqueous solution, primary interest centers on the production of CO, formic acid, methanol and alcohols, and methane and hydrocarbons. The standard redox potentials (versus the saturated calomel electrode, SCE) for the common CO2 reduction products of formic acid, CO, formaldehyde, methanol, and methane in aqueous solution at pH 7.0 are given as [42]:

297

298

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

CO2 + 2 H+ + 2 e − ↔ HCOOH

(E ° = 0.85 V )

(11.1)

−

CO2 + 2 H + 2 e ↔ CO + H2O

(E ° = −0.76 V )

(11.2)

2 H+ + 2 e − ↔ H2

(E ° = −0.65 V )

(11.3)

CO2 + 4 H + 4 e ↔ HCHO + H2O

(E ° = −0.72 V )

(11.4)

CO2 + 6 H+ + 6 e − ↔ CH3OH + H2O (E ° = −0.62 V )

(11.5)

(E ° = −0.48 V )

(11.6)

+

+

−

CO2 + 8 H + 8 e ↔ CH4 + 2 H2O +

−

The similarity of the reduction potentials for H+ to hydrogen, and for CO2 to various products, explains the issue of competing hydrogen evolution in aqueous solution. Thus, there has been much interest in high hydrogen overpotential metals that provide slow kinetics for H+ reduction. For the ﬁrst electron transfer there are two possible reduction products: the CO2 radical anion, iCO2− ; and the radical •COOH. The latter radical can be formed either directly or by insertion into a metal–hydride bond (as discussed in Section 11.2.1). In aqueous solution, the production of HCOOH and CO is thought to proceed primarily through the •COOH species, the dehydration of which , through an additional electron transfer coupled to protonation, can lead to CO and H2O. Similarly, a proton-coupled electron transfer can lead to HCOOH. Various theoretical factors leading to the production of either CO or HCOOH at different metals has been discussed by Frese, including the adsorption properties of the species on the electrode surface [41]. Similarly, CO2 reduction mechanisms at metal electrodes have been discussed by Taniguchi [44]. Further reduction to HCHO, CH3OH, and CH4 is less understood, although the spectroscopic identiﬁcation of some intermediates has shed light on the reduction mechanisms. The production of HCHO and CH3OH is thought to proceed through an ·HC O intermediate, while the production of CH4 is assumed to proceed through CO reduction, ﬁrst to adsorbed C, then forming a surface-adsorbed carbene. A more detailed mechanistic analysis of this process has been created by Frese [41]. 11.3.1.1 Reduction to CO and HCOOH In the very early studies conducted by Coehn and Jahn, an aqueous reduction of CO2 to formic acid was achieved using zinc amalgam electrodes, and resulting in an 89% faradaic efﬁciency for formic acid [45]. Others were able to increase these yields to almost 100% faradaic efﬁciency by changing the electrolyte and cathode preparations [46]. Since then, there has been much interest in CO2 reduction to the two-electron reduced species of formic acid. For example, Ryu et al. obtained a 100% faradaic efﬁciency for formate production at a Hg electrode in neutral aqueous solution [47], with Tafel plots showing two different regions relating to a consecutive two-electron step mechanism. Overall, the current densities were fairly large for these studies, although high overpotentials were an issue. Similarly, the kinetics and mechanism of CO2 reduction to formic acid has been studied by Vassiliev et al. over a wide range of metal electrodes with high or moderate hydrogen overpotentials (e.g., Cu, Sn, In, Sb, Bi, In, Zn, Pb) [48]. The best

11.3 Electrochemical Reduction

electrode materials were those that showed maximal adsorption of CO2. Experimental evidence, similar to results achieved with Hg as presented above, suggested that the ﬁrst and second electrons were transferred to a surface-adsorbed species of CO2(ads) and, then to the radical anion, iCO2− (ads), to produce formate through reaction with either water or H3O+, depending on the solution pH. Hori et al. also studied many metal electrodes in neutral and mildly acidic solutions [49]. In 0.5 M KHCO3 Cd, Sn, Pb, and In were observed to primarily produce formic acid, with faradaic efﬁciencies of 66%, 73%, 81%, and 95%, respectively. The current densities were held at 5.5 mA cm−2, and potentials ranged from −1.6 to −1.89 V (versus SCE). Both, Ag and Au were observed to produce primarily CO, with faradaic efﬁciencies as high as 90% and 87%, and at potentials of −1.69 and −1.39 V (versus SCE), respectively. Neither Fe nor Ni were observed to reduce CO2 to any products; rather, H2 production dominated in these cases. Others have proposed that Ni is not catalytic for CO2 reduction due to a very strong adsorption of the intermediates and CO [50, 51]. The reduction of CO2 at Zn electrodes was not consistent, with faradaic efﬁciencies for formic acid and CO ranging from 17% to 85%, and from 3% to 63%, respectively [49]. Frese reported that Co produced CO with faradaic efﬁciencies ranging from 6% to 26% depending on potential. Similarly, under the correct conditions, it was found that Fe and Pd could produce CO in faradaic yields of 29% and 31%, respectively [52–54]. Using a Pd wire, Azuma et al. reduced CO in CO2-saturated aqueous solutions of 0.05 M KHCO3 at −2.0 V (versus SCE), but the faradaic yields for formic acid and CO were only 16% and 12%, respectively, with mainly H2 formed [55]. Reduction at Pt electrodes and Pt group metals of Ir and Rh in aqueous solution has led to CO production [56–58]. Intermediate-adsorbed species have been studied by Giner in acidic solution at Pt [59], while similar spectroscopic experiments at Pt electrodes have also been conducted by Nicolic et al. [60]. It appears that the reduction of CO2 occurs by reaction with adsorbed hydrogen atoms to produce both linear and bridge-bonded CO. The reduced CO2 species at Pt electrodes, however, caused a rapid poisoning of the electrodes to further CO2 reduction [61]. 11.3.1.2 Reduction to CH3OH and Alcohols There has been much interest in reducing CO2 to methanol for use in fuel cells. In one of the earliest reports, when Summers et al. used molybdenum electrodes in CO2-saturated acidic solutions (pH 4.2.); at −0.7 to −0.8 V (versus SCE), methanol was formed with faradaic efﬁciencies approaching 100% [62]. Such efﬁciencies were observed to increase to 370% when the electrode was cycled at −1.2 to +0.2 V (versus SCE), indicating that the electrode was being corroded and becoming unstable. Reports have also been made on the reduction of CO2 to methanol at hydrogenated palladium electrodes in an aqueous 0.5 M NaClO4 electrolyte at pH 5.2 with 10 mM pyridine [63]. For this, the Pd electrodes were preloaded with hydrogen by hydrogenation in 1 M H2SO4 at 1–3 mA cm−2, with pyridinium serving as the electrocatalyst. The electrolyses, when carried out at 0.05 mA cm−2, gave faradaic efﬁciencies for methanol of up to 30%. The potential was only −0.7 V (versus SCE),

299

300

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

and the onset for CO2 reduction was similar to the thermodynamic potential for CO2 reduction to methanol. Labeling studies with 13CO2 veriﬁed the formation of 13 CH3OH directly from CO2. Likewise, in the absence of pyridine, methanol was not produced; the substitution of pyridine with N-methylpyridine did not produce methanol, which indicated that the N–H hydrogen atom of pyridinium was indeed critical for the reduction of CO2. Interestingly, the system appeared to be stable for over 60 h of continual electrolysis, unlike most of the electrochemical systems used for CO2 reduction. In an aqueous CO2-saturated Na2SO4 electrolyte, using electroplated Ru electrodes, Frese and Leach observed faradaic efﬁciencies of up to 42% for methanol production at a temperature of 333 K at a potential of only −0.55 V (versus SCE) [54]. Faradaic yields of up to 30% were likewise obtained for methane. When Popic et al. examined RuO2 electrodes, either alone or with Cu and Cd adatoms [64], in 0.5 M NaHCO3 at a potential of −0.8 V (versus SCE), they were able to reduce CO2 to methanol with faradaic efﬁciencies of 17%, 41%, and 38% after 480 min of electrolysis for RuO2, RuO2/Cu, and RuO2/Cd electrodes, respectively. When Bandi and Kuhne studied the reduction of CO2 to methanol at mixed RuO2 + TiO2 electrodes (ratio 3 : 1) produced by coating titanium foil [65], in a CO2-saturated KHCO3 solution at a current density of 5 mA cm−2, only minimal CO2 reduction was observed. However, the addition of electrodeposited Cu led to faradaic efﬁciencies of up to 30% for methanol at potentials of approximately −0.972 V (versus SCE). Trace amounts of formic acid and ethanol were also observed. In this case, the rate-limiting step was surmised to be the surface recombination of adsorbed hydrogen and iCO2− to yield adsorbed COOH−. More recently, Qu et al. examined composite RuO2/TiO2 nanotube and nanoparticle platinum electrodes [66]. In a 0.5 M NaHCO3 solution at −0.8 V (versus SCE), the nanoparticle-based electrodes yielded faradaic efﬁciencies for methanol of 40%, compared to 61% for the nanotube composites However, no explanation was offered as to why the nanotube-based electrodes provided an increased catalytic activity. In general, few electrochemical systems have been shown to reduce CO2 to higher-order alcohols, as most operate at fairly large overpotentials and produce low faradaic yields and selectivity. Details of these systems are provided elsewhere [42]. 11.3.1.3 Reduction to CH4 and Hydrocarbons Most reports of CO2 reduction to methane and hydrocarbons have involved Cu electrode materials. Thus, Cu electrochemistry in relation to CO2 reduction has been extensively studied, and the ﬁeld widely reviewed [67]. Extensive information on CO2 reduction to methane is available in several excellent reviews [41, 42, 67]. Hori et al. was one of the ﬁrst to report on the reduction of CO2 to methane in aqueous solution [49, 68]. Here, at a current density of 5 mA cm−2 in 0.5 M KHCO3 electrolyte, a temperature dependence of faradaic efﬁciency was observed for the observed products such that, at 273 K, the efﬁciency for methane was 70%. Moreover, the efﬁciency was shown to decrease linearly as the temperature was raised,

11.3 Electrochemical Reduction

reaching 0% at 316 K. By comparison, the productions of hydrogen, CO, and ethylene were each observed to increase with increasing temperature. Ikeda et al. observed similar faradaic efﬁciencies at Cu-coated glassy carbon and Pt electrodes [69]. The reason for increased methane yields at lower temperatures was attributed to the presence of longer-lived intermediates. Other experiments showed that both selectivity and product distribution could be drastically affected by changes in the electrode potentials [70]. Similarly, surface roughness and pretreatment were also factors affecting methane production [41]. In addition, when the effect of crystal structure was examined, the rate of methane production in a CO2-saturated 0.5 M KHCO3 electrolyte was shown to be highest on Cu(111), followed by Cu(110) and Cu(100) [41]. While the overpotentials are high for methane production, rather high current densities have also been achieved at Cu electrodes. For example, Cook et al. were able to reduce CO2 to methane in aqueous 0.5 M KHCO3 solutions at a current density of 38 mA cm−2, with 33% faradaic efﬁciency [71], although the potential was −2.29 V (versus SCE). Subsequently, it proved possible to increase the faradaic yields to 79% for methane and ethene together on Cu-coated glassy carbon electrodes from in situ Cu deposition, but in this case the potential was −2.0 V (versus SCE), in the same electrolyte with current densities of up to 25 mA cm−2. The mechanism of CO2 reduction to methane at Cu electrodes has been proposed by various groups [72–74], most of which involved the splitting of adsorbed CO followed by the hydrogenation of surface C atoms. When DeWulf et al. used X-ray photoelectron spectroscopy (XPS) and Auger electron spectroscopy to study the reaction [72], they observed surface-bound carbenes (Cu CH2) as an intermediate in the system. Likewise, others used both in situ infrared (IR) reﬂection absorption spectroscopy and surface-enhanced Raman spectroscopy to observe the initial product of CO2 reduction on Cu [74]. Typically, two different linearly bound CO species were identiﬁed and attributed to adsorption on either surface defect sites or terraces. Whilst copper electrodes are capable of catalyzing the reduction of CO2 to methane, the electrodes normally become deactivated within a short time due to carbon deposits and copper-oxide species. Consequently, various attempts have been made, using pulsed electroreduction, to prevent this deactivation. When Shiratsuchi et al. and Nogami et al. each used pulses of cathodic and anodic biases between 0 and −1.8 V (versus SCE) [75, 76], the anodic pulses were observed to reoxidize the carbon deposits on the electrode. However, in CO-saturated KHCO3 solution, the faradaic efﬁciencies for methane and ethylene remained constant at about 10% and 25%, respectively. The best results were achieved at 10 °C, using a cathodic bias of −2.6 V (versus SCE), which yielded a total faradaic efﬁciency of 65% for methane + ethylene. In a similar approach, Jermann and Augustynski used the pulse technique in aqueous CO2-saturated 0.5 M KHCO3 solutions [77]; in this case, pulsing between −1.96 V (versus SCE) to +1.05 V every 5 min led to faradaic efﬁciencies of 46% and 6% for methane and ethylene, respectively. Frese et al. also studied the reduction of CO2 to methane in aqueous Na2SO4 solutions at Ru electrodes [54]. The reaction rate was observed to be sensitive to

301

302

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

pH, temperature, electrolyte purity, and hydrogen atom and CO coverage. Frese has discussed these issues in detail [41]. Unlike other systems for methane production, a minimal overpotential was needed; in fact, at a potential of −0.545 V (versus SCE) at 353 K and a pH of 2.7, a faradaic efﬁciency in excess of 27% for CH4 was achieved. However, the current densities approached only ∼0.4 mA cm−2, and instability was an issue. 11.3.2 Reduction in Nonaqueous Solutions at Metal Electrodes

In aprotic organic solvents, where there are no protons available to be involved in CO2 reduction, the primary products are CO and oxalate, and all reductions must proceed through the CO2 radical anion, iCO2− , as discussed previously. The relevant reactions in aprotic media are listed below, as summarized by Halmann and Steinberg [42]. CO2 + e − ↔ iCO2−

(E ° = −2.21 V versus SCE)

(11.7)

2 iCO2− ↔ C2O4 2−

(oxalate production )

(11.8)

2 iCO → O = C − O (C = O) − O − 2

−

O = C − O (C = O) − O− ↔ CO32− + CO (CO production )

(11.9) (11.10)

Eggins and McNeill compared the solvents of water, dimethylsulfoxide (DMSO), acetonitrile, propylene carbonate, and DMF electrolytes for CO2 reduction at glassy carbon, Hg, Pt, Au, and Pb electrodes [78]. The main products were CO and oxalate in the organic solvents, while metal electrodes (such as Pt) which absorb CO2 showed a higher production for CO. In DMF, containing 0.1 M tetrabutyl ammonium perchlorate and 0.02 M CO2 at a Hg electrode, Isse et al. produced oxalate and CO with faradaic efﬁciencies of 84% and 1.7%, respectively [79]. Similarly, Ito et al. examined a survey of metals for CO2 reduction in nonaqueous solution, and found that Hg, Tl, and Pb yielded primarily oxalate, while Cu, Zn, In, Sn, and Au gave CO [80, 81]. Kaiser and Heitz examined Hg and steel (Cr/Ni/Mo, 18 : 10 : 2%) electrodes to produce oxalate with 61% faradaic efﬁciency at 6 mA cm−2 [82]. For this, they examined the reduction of CO2 at electrodes where CO2 and reduction products do not readily adsorb. The production of oxalate was therefore explained by a high concentration of CO2 radical anions, iCO2− , close to the surface. Dimerization resulted in oxalate production rather than CO formation. When Desilvestro and Pons used in situ IR reﬂection spectroelectrochemistry to observe the reduction of CO2 to oxalate at Pt electrodes in acetonitrile [83], two different forms of oxalate were observed. Similarly, Aylmer-Kelly et al. studied CO2 reduction in acetonitrile and propylene carbonate at Pb electrodes [84], by using modulated specular electroreﬂectance spectroscopy. Subsequently, two radical intermediates were observed which they determined to be the CO2 radical anion, iCO2− , and the product of the radical anion and CO2, the (CO2 )2− adduct (see Equations 11.9 and 11.10). Vassiliev et al. also studied the reduction of CO2 in

11.3 Electrochemical Reduction

nonaqueous media at Hg, Pb, Sn, In, and Pt electrodes [61], by examining the overpotentials for CO2 reduction at each metal and the corresponding Tafel slopes. The sensitivity of overpotential with electrode materials suggested that the ratelimiting step involved an absorbed species. On all electrodes other than Pt, two Tafel regions were observed. The ﬁrst region, a low overpotential Tafel region, with a slope of 140–180 mV, was proposed to correspond to the reaction of the radical anion, •CO2, with CO2 to form the (CO2 )2− adduct that would ultimately yield oxalate. The second region, a high overpotential Tafel region, was thought to correspond to the initial ﬁrst electron reduction of CO2 to the radical anion, •CO2. On Pt, diproportionation was proposed as the mechanism for CO formation, as observed by Sawyer (see below). Haynes and Sawyer observed CO2 reduction in DMSO at Au electrodes. In this case, CO was produced, with the mechanism proceeding through a disproportionation of two CO2 radical anions to form CO and CO32− [85]. 11.3.3 Reduction Mediated by Metal Complexes

Due to the high overpotentials required at metal electrodes for the reduction of CO2, many groups have studied a variety of metal complexes as electron mediators that can be reduced at lower potentials. In general, however, the reported metal complex systems for CO2 reduction still operate with substantial overpotentials. In aqueous solution, the reduction product is usually formic acid (as seen in Section 11.2). Similarly, reduction in nonaqueous solution leads to CO and oxalate formation. The mechanistic pathways for CO2 reduction relate to the descriptions provided in Section 11.2, and proceed through intermediate metal–CO2 complexes or by the direct reduction of CO2 through an outer-sphere electron transfer. Keene and Sullivan have reviewed the mechanisms of CO2 reduction by metal complexes in detail [9]. Benson et al. and Halmann and Steinberg have recently reviewed the topic of homogenous electrocatalysis [42, 86], while Halmann and Steinberg have described metal complexes embedded into polymer coatings on electrode surfaces. It is unclear whether this approach is better than the use of solution-based catalysts, and embedded catalysts appear to be more stable than their solution-based counterparts, although the polymer matrix typically breaks down leading to deactivation by other means. At this point, therefore, only those key ﬁndings regarding solution-based electron mediators for CO2 reduction will be discussed. 11.3.3.1 Transition Metal Complexes The rhenium complexes described in Section 11.2 have also been studied as electron mediators for CO2 reduction at metal electrodes. Hawecker et al. used the complex Re(bpy)(CO)3Cl in DMF/water (9 : 1) at glassy carbon electrodes at a potential of −1.44 V (versus SCE) to produce CO with 98% faradaic efﬁciency [15, 87]. Likewise, Sullivan et al. reported the production of CO with similar efﬁciency at a platinum electrode at −1.5 V (versus SCE) by using the complex fac-Re(bpy) (CO)3Cl [88]. Ruthenium complexes that have been used in photochemical

303

304

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

catalysts for CO2 reduction have also been studied as electron mediators in electrochemical systems. For example, Ishida et al. studied [Ru(bpy)2(CO)2]2+ and [Ru(bpy)2(CO)Cl]+ at Hg electrodes in either pH 6 or pH 9.5 CO2-saturated water/ DMF (9 : 1, v/v) [89–91]. In this case, at −1.5 V (versus SCE), CO and H2 were produced at pH 6, and formate at pH 9.5, with 34% faradaic efﬁciency. The mechanism was proposed to proceed via a two-electron reduced Ru(0) intermediate, which coordinated the CO2. Protonation, followed by hydroxide loss, led to the production of CO, while no hydroxide loss led to formate production. The formate production was also increased to upwards of 84% faradaic efﬁciency in acetonitrile solution, with methyl-amines or phenol serving as the proton sources. Various other metal complexes have also been shown to catalyze CO2 reduction. For example, Bruce et al. were able to produce CO with 90% faradaic efﬁciency at −1.4 to −1.6 V (versus SCE) at a Pt mesh electrode using the complex cis[Os(bpy)2(CO)H][PF6] in anhydrous acetonitrile [92, 93]. By adding water, it was possible to produce formic acid with up to 25% faradaic efﬁciency. The mechanism was considered to proceed via the di-reduced complex, which coordinated CO2. In a similar approach, Bolinger et al. studied cis-[Rh(III)(bpy)2Cl2]2+ in acetonitrile using a carbon cloth electrode at −1.56 V (versus SCE) [94], and were able to reduce CO2 to formic acid with 80% faradaic efﬁciency. The protons involved in this reaction were assumed to derive from the supporting electrolyte, [(n-Bu)4N](PF6). DuBois and Miedaner also synthesized a variety of Pd complex CO2 reduction catalysts of the composition [Pd(triphosphine)L](BF4)2, where triphosphine is a triphosphine ligand and L = CH3N, P(OMe)3, PEt3, P(CH2OH)3, and PPh3 [95]. As a result, maximum efﬁciencies of up to 85% were reported for CO formation in acidic acetonitrile. The E1/2 values of many such complexes were very close to the thermodynamic reduction potentials, with only 100–200 mV over the respective E1/2 being required for bulk electrolyses. A subsequent mechanistic analysis proposed the existence of a Pd(I) species that interacted with CO2 [96]. 11.3.3.2 Macrocyclic Complexes Fisher and Eisenberg showed that Ni(II) and Co(II) tetra-azamacrocyclic complexes in water–acetonitrile mixtures at Hg electrodes reduced CO2 to CO, H2, and formic acid at −1.5 V (versus SCE) [97]. Here, the total faradaic efﬁciencies approached 100%, but the ratios for CO : H2 were 1 : 1. Tinnemans et al. also investigated several Ni and Co tetra-azamacrocyclic complexes as electron mediators at Hg electrodes [33]. As an example, in DMF with 5% water at −1.3 V (versus SCE), up to 66% faradaic efﬁciency was achieved for CO with the complex {Co(II)[Me2Pyo(14)trieneN4]}2+, using Hg electrodes. Beley et al. and others studied tetra-azamacrocyclic Ni complexes, similar to those presented in Section 11.2, in aqueous and organic solvents for the mediated reduction of CO2 [98–100]. In aqueous 0.1 M KNO3 solution at a potential of −1.2 V (versus SCE), Ni(II)-cyclam dichloride (cyclam = 1,4,8,11-tetra-azatetradecane) reduced CO2 to CO with 96% faradaic efﬁciency at Hg electrodes. The mechanism involved a ﬁrst electron reduction of the species which coordinated CO2, followed by CO2 protonation, and a second electron transfer to yield CO and OH− (as dis-

11.4 Semiconductor Systems for Reduction

cussed in Section 11.2), thereby regenerating the Ni(II) complex [12]. Balazs and Anson showed that the system deactivation was caused by the deposition of a Ni(0) complex, cyclam, and CO onto the electrode surface [101]. More recently, Rudolph et al. was able to reduce CO2 to oxalate with faradaic efﬁciencies approaching 100% with their most active and stable complex [102]. These authors examined a variety of macrocyclic nickel chelate complexes with various substituent groups on the ring in acetonitrile solution. Whilst it is interesting that the group was able to produce oxalate catalyzed by a metal complex, the potentials required for reduction were −1.9 to −2.2 V (versus SCE), similar to the potential required for the direct reduction of CO2 in aprotic solvent (−2.21 V versus SCE). The very negative potentials in this reaction highlight the overall theme of the electrochemical reduction of CO2. 11.3.3.3 Metal-Containing Enzyme-Mediating Complexes Recently, interest has been expressed in natural enzymes that effect the reduction of CO2 to various products (see Section 11.4). For example, Reda et al. used a tungsten-containing formate dehydrogenase 1 enzyme derived from Syntrophobacter fumaroxidans in the mediated electroreduction of CO2 to formate [103]. The enzyme, which is either adsorbed onto the graphite electrode surface or is free in solution, was observed to reduce CO2 to formate with near-100% faradaic efﬁciency. Although a minimal overpotential for the process was required (∼0.4 V of applied bias), the current densities were rather low.

11.4 Semiconductor Systems for Reduction

Similar to the molecular photosensitizers described above, solid semiconductor materials can absorb photons and convert light into electrical energy capable of reducing CO2. In solution, a semiconductor will absorb light, and the electric ﬁeld created at the solid–liquid interface effects the separation of photo-excited electronhole pairs. The electrons can then carry out an interfacial reduction reaction at one site, while the holes can perform an interfacial oxidation at a separate site. In the following sections, details will be provided of the reduction of CO2 at both bulk semiconductor electrodes that resemble their metal electrode counterparts, and semiconductor powders and colloids that approach the molecular length scale. Further information on semiconductor systems for CO2 reduction is available in several excellent reviews [8, 44, 104, 105]. 11.4.1 Photoelectrochemical Semiconductor Electrode Systems

In bulk semiconductor electrodes, the electric ﬁeld at the solid–liquid interface causes the electrons and holes to move in opposite directions within the material. Then, depending on the type of material, n-type or p-type, the electrons or holes

305

306

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

will move either to the electrode surface or into the bulk. If there is a need to carry out a reduction at the illuminated electrode surface, then a p-type semiconductor electrode must be used, in which the electric ﬁeld drives the electrons to the solid– liquid interface and the holes into the bulk. The opposite is observed for illuminated n-type materials, where the illuminated interface produces oxidation products. However, in the dark, n-type semiconductor materials are capable of reducing CO2. In such systems no light is needed, and the material behaves similar to a metal electrode. As with a classic electrochemical cell, a photoelectrochemical cell employs a counterelectrode to complete both the circuit and the catalytic cycle. In the case of a p-type photocathode-based cell, the counterelectrode uses the holes generated by the photoinduced charge separation to complete an oxidation reaction. Typically, the counterelectrode is simply a metal electrode, although some groups have used illuminated n-type semiconductor materials capable of photo-oxidations. At this point, CO2 reduction is reviewed for both aqueous and nonaqueous photoelectrochemical systems. In all systems, the main issue continues to be the high overpotentials required to reduce CO2, which means that electrical energy is needed in addition to light. As described for metal electrodes, the competition of hydrogen evolution at photocathodes also continues to be a drawback in aqueous solution. Similarly, the stability of semiconductor materials in water has long been an issue, especially for n-type semiconductors used for photo-oxidations. The p-type materials described here – and especially the larger bandgap materials – are fairly stable in aqueous media. However, the smaller bandgap p-type semiconductor electrodes that match the solar spectrum have been reported to be rather unstable [8]. Thus, the use of surface-modiﬁed semiconductor electrodes as a means to increase stability will be examined. 11.4.1.1 Unmodiﬁed Semiconductor Electrode Interfaces Halmann was the ﬁrst to report the reduction of CO2 at an illuminated p-GaP semiconductor electrode [104]. In an aqueous solution buffered at pH 6.8, illuminated with 365 nm light, mostly formic acid (plus smaller amounts of formaldehyde and methanol) was produced, although the electrode was biased at −1.0 V (versus SCE). Halmann reported a faradaic efﬁciency of 60% for methanol at a potential of −1.4 V (versus SCE) when using a TiO2 semiconductor counterelectrode. In similar experiments conducted by Inoue et al., formic acid, formaldehyde, and methanol were also observed using a p-GaP electrode, at a potential of −1.5 V (versus SCE) in aqueous 0.5 M H2SO4 [106]. Again, however, high overpotentials meant that no optical energy was being converted and stored in the products. One major issue with p-GaP is its large bandgap, which allows only 17% of the solar spectrum to be absorbed. When a material with a smaller bandgap, namely p-CdTe, was examined for CO2 reduction, Taniguchi et al. obtained faradaic efﬁciencies of 70% for CO formation in DMF with 5% water [107]. For this, the electrode was illuminated with visible light at 600 nm with tetrabutylammonium perchlorate (TBAP) as the electrolyte. Although the quantum yields approached unity, the electrode was biased at −1.6 V (versus SCE), which indicated a require-

11.4 Semiconductor Systems for Reduction

ment for electrical energy as well as light. In other experiments conducted by Taniguchi et al. and Yoneyama et al., the presence of tetra-alkylammonium salts was observed to be critical for the obtained current efﬁciencies [108, 109]. Canﬁeld and Frese also investigated smaller-bandgap semiconductor electrodes [110] when, in aqueous solutions at pH 4, faradaic efﬁciencies of 57% and 80% were observed for methanol production at p-GaAs and p-InP electrodes, respectively. Trace amounts of formaldehyde were also observed. The electrodes were biased at −1.2 to −1.4 V (versus SCE) and illuminated using a tungsten-halogen lamp. Frese and Canﬁeld also observed 100% current efﬁciency for the production of methanol at n-GaAs in the same solution, and at a potential of −1.2 to −1.4 V (versus SCE) [111]. Unfortunately, these experiments were conducted in the dark and so related more to the electrode materials (as discussed in Section 11.3). In all of these cases, the extreme negative potentials required to effect the reduction meant that little to none of the light energy employed was captured in the products. Rather, these systems represent light-activated electrochemical cells, where the power to carry out the reduction of interest was derived almost exclusively from an external electrical power supply. 11.4.1.2 Modiﬁed Semiconductor Electrodes In order to increase the rates of CO2 reduction at semiconductor electrodes, various groups have modiﬁed the electrode surfaces with catalytic agents that effect CO2 reduction, without blocking light absorption. As an example, Ikeda et al. examined the effect of electroplating and sputtering of Pb, Zn, and Au onto p-GaP [112]. In aqueous solution at −1.25 V (versus SCE), both Pb and Zn were observed to increase the current efﬁciency for CO2 reduction to CO and formic acid, but Au showed little effect. The addition of these high hydrogen overpotential metals was thought to increase CO2 reduction by slowing the reduction of H+. Similarly, Ito and colleagues were able to produce methane with 7% faradaic efﬁciency using Cu-coated p-GaP electrodes at a potential of −1.6 V (versus SCE) [113]. Hinogami et al. deposited small metal particles of Cu, Ag, and Au (20–200 nm diameter) onto p-Si in order to reduce the overpotential for CO2 reduction [114], such that the onset potentials were ∼0.5 V more positive than those of the respective bulk metal electrode. In a CO2-saturated aqueous solution, mainly CO was produced with some formic acid; however, at −1.05 V (versus SCE) the Ag-coated electrodes produced CO with 51% faradaic efﬁciency. Similarly, Au-coated electrodes yielded CO with 62% faradaic efﬁciency at −0.74 V (versus SCE). The Cucoated electrodes, when held at a potential of −1.2 V (versus SCE), produced formic acid and CO with faradaic efﬁciencies of 32% and 19%, respectively. Small amounts of methane and ethylene were also observed. More recently, Kaneco et al. evaluated Pb, Ag, Au, Pd, Cu, and Ni sputter-coated p-InP electrodes [115]. Whilst the onset potentials for CO2 reduction were less negative than those observed at the corresponding metal electrodes, the authors chose an extremely negative potential of −2.55 V (versus SCE) for their studies. By using a LiOH/methanol electrolyte, they examined the dependence of the amount of metal deposited on the current efﬁciency of CO2 reduction, where the main

307

308

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

products were CO and formic acid. The maximum faradaic yield for CO was 80% at an Ag-modiﬁed p-InP; however, the Pd-modiﬁed electrodes were selective for CO, with 55% faradaic efﬁciency. Many of the earliest studies focused on the use of polymer-coated semiconductor materials for the reduction of CO2. An example was the study of Aurian-Blajeni et al., who electropolymerized polyaniline onto p-Si [116]. In an aqueous CO2saturated solution, a total faradaic efﬁciency for formic acid and formaldehyde of 28% was achieved, but at a potential of −1.9 V (versus SCE). Likewise, Cabrera and Abruna electropolymerized [Re(CO)3(v-bpy)Cl], where v-bpy is 4-vinyl-4′methyl-2,2′-bipyridine [117]. For CO production, TONs of about 450 were observed, while the faradaic efﬁciencies approached 100%. Upon illumination in acetonitrile solution, the onset potential for reduction was −0.65 V (versus SCE). Daube et al. also reported the reduction of CO2 at redox polymer-coated p-Si electrode containing a Pd catalyst [118]. In an aqueous bicarbonate solution, formic acid was produced with 70% faradaic yield and a small overpotential. 11.4.1.3 Homogenous Solution Catalysts at Semiconductor Electrodes Another method of stabilizing the surface of semiconductor electrodes relates to an electrolyte modiﬁcation, using solution mediators that efﬁciently accept the electron from the semiconductor to subsequently reduce CO2. In aqueous solution, Taniguchi et al. used a p-GaP photoelectrode in the presence of 15-crown-5 ether at a potential of −0.95 V (versus SCE) [119]. In this case, current efﬁciencies of 44%, 15%, and 4% were observed for methanol, formic acid, and formaldehyde, respectively. Similarly, when Bockris and Wass used various organic mediators in a DMF solution with 5% water and p-CdTe as the photocathode [120], the bare p-CdTe produced a faradaic efﬁciency for CO of 92%. Subsequently, by using their best catalysts, namely 15-crown-5 ether and 18-crown-6 ether, these authors were able also to produce methanol with faradaic efﬁciencies of 14% and 13%, respectively, with the remaining current going to CO production. Although the potential at which these electrolyses were carried out was not reported, the onset potential for CO2 reduction was shifted some 400 mV anodically in the presence of the organic mediators. Both, Beley et al. and Petit et al. reported the reduction of CO2 to CO using Ni(cyclam)2+ catalysts in aqueous solution at p-GaAs and p-GaP photoelectrodes [121–123]. Here, the faradaic yields for CO approached 100%, at potentials of −1.0 V and −0.44 V (versus SCE), respectively, although carbon deposits on the surfaces of the electrodes led eventually to a degradation of the system. In order to create a highly selective system that operated at minimal overpotential, Parkinson and Weaver used a p-InP photoelectrode, the electron mediator methyl viologen, and the enzyme formate dehydrogenase to reduce CO2 in aqueous solution to formate [124]. Although faradaic efﬁciencies of 93% were achieved at an applied voltage of only −0.2 V (versus SCE), the stability of the system proved to be a major issue, with trace amounts of oxygen leading to decomposition of the enzyme and methyl viologen.

11.4 Semiconductor Systems for Reduction

More recently, the use of a pyridinium mediator in an aqueous p-GaP photoelectrochemical system illuminated with 365 nm and 465 nm light has been reported [125]. In this case, a near-100% faradaic efﬁciency was obtained for methanol production at underpotentials of 300–500 mV from the thermodynamic CO2/methanol couple. Moreover, quantum efﬁciencies of up to 44% were obtained. The most important point here, however, was that this was the ﬁrst report of CO2 reduction in a photoelectrochemical system that required no input of external electrical energy, with the reduction of CO2 being effected solely by incident light energy. 11.4.2 Heterogeneous Photochemical Semiconductor Systems

Heterogeneous semiconductor systems involve either suspensions or slurries of larger-sized semiconductor powders, or smaller colloids in solution. In principle, these semiconductor particles may act as tiny photoelectrolysis cells, similar to the photoelectrochemical systems discussed above. However, as many of the materials used for bulk electrodes are also described here in particulate form, both similar and new problems may arise, most notably irreproducibility in particle preparation, stability issues, and low CO2 reduction rates. In heterogeneous semiconductor systems, the entire electrochemical cycle occurs on the surface of the particles. Typically, photogenerated electrons reduce CO2, while the corresponding holes complete the catalytic cycle by performing an oxidation reaction. Since it is impossible to separate the reduction and oxidation reactions, a hole scavenger (i.e., a sacriﬁcial reductant) is typically added to minimize the reoxidation of reduced CO2 species, and is itself oxidized. When in their powder form, semiconductors have been ground such that their particle sizes are on the order of microns; by comparison, colloidal particles typically range in size from a few to some hundreds of nanometers. Unlike larger semiconductor powder suspensions, colloids are evenly dispersed mixtures that have an homogeneous appearance. However, due to their small particle size they have been found to have different light absorption and chemical reactivity properties when compared to larger particles. Rossetti et al. explained that quantum effects in these small particles cause conﬁnement of the charge carriers, which leads to perturbations in the semiconductor’s band structure [126] that may, in turn, cause increases in the effective bandgap. This results in enhanced reduction and oxidation potentials for the electrons and holes, a phenomenon that would be expected to enhance the thermodynamics of CO2 reduction at the colloid interface. Unfortunately, however, the need to use higher-energy photons bears associated costs. 11.4.2.1 Unmodiﬁed Semiconductor Colloids and Powders Among the earliest reports of semiconductor powders being used for the reduction of CO2 was one made by Inoue et al. [106], who examined a wide range of semiconductors (WO3, TiO2, ZnO, CdS, GaP, SiC, in 200–400 mesh), illuminated in aqueous solution. The study results showed that CO2 could indeed be reduced to

309

310

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

formate, formaldehyde, and methanol, and that the product yields correlated qualitatively to the position of the conduction band edge of the semiconductors. In the case of WO3, which had a conduction band edge that was insufﬁciently high in energy to reduce CO2, no products were generated; yet, in contrast, SiC gave the highest yields as it had the most negative conduction band energy. When Yamamura et al. studied 100- and 1000-mesh powders of SiC, the former particles yielded only trace amounts of products, while the latter produced both methanol and ethanol [127]. High quantum yields of 80% were achieved by Henglein et al. for the reduction of CO2 to formate using ZnS colloidal suspensions illuminated with ultraviolet light [128]. Here, the electrolyte was a mixture of water and ethanol, and 2-propanol acted as the hole scavenger. With colloidal ZnS in CO2-saturated aqueous solutions containing tetramethyl ammonium salts under illumination by an arc-lamp, Eggins et al. observed the production of formic acid and formaldehyde along with two-carbon (oxalic, glycolic, and glyoxylic) and even four-carbon (tartaric) acids [129]. ZnS colloids were also used by Kuwabata et al. to photoreduce CO2 to formate [130]. In this system, the ZnS colloids also reduced pyrroloquinoline quinone which served as an electron mediator to the enzyme, methanol dehydrogenase, which could then reduce formate to methanol. In CO2-saturated aqueous solution at pH 7 and under far-UV (280 nm) illumination, quantum efﬁciencies of 7% and 6% were achieved for formate and methanol, respectively. Colloidal particles of CdS were examined by Eggins et al. [131]. In CO2-saturated aqueous solutions containing tetramethyl ammonium chloride, illumination with an arc-lamp produced formic acid and formaldehyde along with glyoxylic acid and acetic acid. Either hydroquinone or Na2SO3 was used as the hole scavenger to increase yields. When CdSe colloids were examined for CO2 photoreduction, those with a particle size <50 Å were seen to produce formic acid under illumination in aqueous solution, whilst larger particles exhibited no photoreduction [132]. 11.4.2.2 Metal-Coated Semiconductor Colloids and Powders In another study conducted by Yamamura et al. [133], 1000-mesh SiC powders were ﬁrst coated with Pb, Cu, Pd, Fe, or Pt, after which irradiation of their aqueous suspensions with a xenon lamp led to the production of formic acid, formaldehyde, methanol, acetaldehyde, and ethanol. The photoyields were very small, however. Similarly, Cook et al. either coated p-SiC powders with Cu, or added Cu powder to the semiconductor suspensions [134]. Interestingly, the Cu-coated SiC powders produced methane, although the rate of production deteriorated over time. The p-SiC and Cu powder mixture, however, appeared to provide a degree of stability for methane production. Whilst many studies of metal-coated TiO2 powders and colloids for the reduction of CO2 have been reported [135], TiO2 alone seems ineffective for CO2 reduction. Adachi et al. was able to reduce CO2 to methane, ethane, and ethylene in aqueous solution using Cu-loaded TiO2 powders [136]. Likewise, others were able to produce methanol almost selectively with both Rh- and WO3-doped TiO2 powders. The

11.5 Concluding Remarks and Future Directions

details of various CO2 reduction systems that involve matrices embedded with TiO2 colloids have been reviewed by Halmann and Steinberg [135], who also discuss the investigations of Anpo and others on gas–solid reactions to convert CO2 to products on various TiO2-anchored catalysts.

11.5 Concluding Remarks and Future Directions

During the past few decades, although many achievements have been made in the ﬁelds of photochemical, electrochemical, and photoelectrochemical reduction of CO2, many key scientiﬁc issues remain to be solved. In all three ﬁelds, the development of inexpensive, stable catalysts that are highly selective for a speciﬁc product and operate at minimal or no overpotential is the main issue. The current densities and TONs of catalysts must also be greatly improved in order to produce economically viable systems. Clearly, catalysis is the key to reducing such overpotentials, to increasing CO2 reduction rates, and to improving selectivity. Future studies should not focus on the speciﬁc approaches discussed here, as these tasks are more fundamental than speciﬁc. As both advantages and disadvantages will be encountered in all three ﬁelds, the ability to combine knowledge from each of these should lead to the creation of an efﬁcient system. Recently, photochemical and photoelectrochemical systems have been cited as the most advantageous in design, mainly because they do not require electrical energy that, invariably, will be derived from fossil fuels. Rather, these systems are simply artiﬁcial photosynthetic schemes that convert light into chemical energy. In the case of photoelectrochemical reduction, however, the systems operate at high overpotentials where electrical energy is required in addition to light. Consequently, the argument that photoelectrochemical systems may be more advantageous than electrochemical systems is lost, as many of the latter operate at low overpotentials and require minimal electrical energy. Hence, an inexpensive, stable metal electrode for selective CO2 reduction operating at low overpotential could be coupled with an efﬁcient photovoltaic device that would supply the necessary electrical energy input. The coupling of an electrochemical system to a photovoltaic device depends on the development of new electrode catalysts, and possibly also new solution catalysts. So, with ongoing problems of competing hydrogen evolution, stability, and low current densities, the quest for abundant and inexpensive materials must continue. The situation is similar for current semiconductor electrodes where, in terms of economic viability, the same problems must be addressed. New semiconductor materials must be developed with bandgaps corresponding to the solar spectrum, and with conduction band energies sufﬁciently energetic for the reduction of CO2. Nonetheless, it is encouraging that various electrochemical and photoelectrochemical systems can reduce CO2 to valuable products, while maintaining their stability. Whilst homogeneous photochemical systems have both advantages and disadvantages, the main issues relate to low reduction rates, to the use of expensive

311

312

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction

metals that is impractical on a large scale, and the need for sacriﬁcial reductants. In many cases, the metal complexes do not serve as true catalysts, but rather are consumed stoichiometrically through a variety of deactivation pathways. Similarly, whilst many reports have been made on the reduction of CO2 to two-electron products such as CO and formate, any further reduction to products such as methanol, methane, and higher-order species is rare. Nonetheless, most systems can efﬁciently capture light, and demonstrate absorption spectra that match the solar spectrum. Indeed, future efforts should be focused on the creation of stable complexes that are selective for highly reduced CO2 products. Likewise, in the catalytic cycle, sacriﬁcial reductants must be replaced with new oxidative catalysts. Whilst many challenges remain with regards to creating economically viable systems capable of converting CO2 into useful materials, such success is rapidly approaching. As not only fuel costs but also concerns about atmospheric CO2 levels continue to rise, these systems may become the key to CO2 mitigation and utilization. The necessary science and technology, though not quite in hand today is within our grasp. An improved understanding of multiple electron reduction processes, coupled to novel designs for new electron mediator catalysts, will surely lead to the successful implementation of a chemical carbon sequestration technology.

References 1 Aresta, M. (ed.) (2003) Carbon Dioxide Recovery and Utilization, Kluwer Academic Publishers, Boston. 2 Aresta, M. and Dibenedetto, A. (2007) Dalton Trans., 2975–2992. 3 Song, C., Gaffney, A.M., and Fujimoto, K. (eds) (2002) CO2 Conversion and Utilization, ACS Symposium Series 809, American Chemical Society, Washington, DC. 4 Royer, E. (1870) C. R. Acad. Sci., 70, 731–735. 5 Fujita, E. (1999) Coord. Chem. Rev., 186, 373–384. 6 Halmann, M.M. and Steinberg, M. (1999) Greenhouse Gas Carbon Dioxide Mitigation: Science and Technology (eds M.M. Halmann and M. Steinberg), Lewis Publishers, Boca Raton, Ch 11. 7 Kirgan, R.A., Sullivan, B.P., and Rillema, D.P. (2007) Top. Curr. Chem., 281, 45–100. 8 Lewis, N.S. and Shreve, G.A. (1993) Electrochemical and Electrocatalytic Reactions of Carbon Dioxide (eds B.P.

9

10

11

12 13

14 15

Sullivan, K. Krist, and H.E. Guard), Elsevier, Amsterdam, Ch 8. Keene, F.R. and Sullivan, B.P. (1993) Electrochemical and Electrocatalytic Reactions of Carbon Dioxide (eds B.P. Sullivan, K. Krist, and H.E. Guard), Elsevier, Amsterdam, Ch 5. Creutz, C. (1993) Electrochemical and Electrocatalytic Reactions of Carbon Dioxide (eds B.P. Sullivan, K. Krist, and H.E. Guard), Elsevier, Amsterdam, Chapter 2. Sutin, N., Creutz, C., and Fujita, E. (1997) Comments Inorg. Chem., 19, 67–92. Collin, J.P. and Sauvage, J.P. (1989) Coord. Chem. Rev., 93, 245–268. Costamagna, J., Ferraudi, G., Canales, J., and Vargas, J. (1996) Coord. Chem. Rev., 148, 221–248. Lehn, J. and Ziessel, R. (1990) J. Organomet. Chem., 382, 157–173. Hawecker, J., Lehn, J.M., and Ziessel, R. (1986) Helv. Chim. Acta, 69, 1990–2012.

References 16 Creutz, C. and Sutin, N. (1985) Coord. Chem. Rev., 64, 321–341. 17 Lehn, J.M. and Ziessel, R. (1982) Proc. Natl Acad. Sci. USA, 79, 701–704. 18 Ziessel, R., Hawecker, J., and Lehn, J.M. (1986) Helv. Chim. Acta, 69, 1065–1084. 19 Hawecker, J., Lehn, J.M., and Ziessel, R. (1983) J. Chem. Soc. Chem. Commun., 536–538. 20 Ishida, H., Tanaka, K., and Tanaka, T. (1988) Chem. Lett., 339–342. 21 Kitamura, N. and Tazuke, S. (1983) Chem. Lett., 1109–1112. 22 Hawecker, J., Lehn, J.M., and Ziessel, R. (1985) J. Chem. Soc. Chem. Commun., 56–58. 23 Kutal, C., Weber, M.A., Ferraudi, G., and Geiger, D. (1985) Organometallics, 4, 2161–2166. 24 Sullivan, B.P. and Meyer, T.J. (1984) J. Chem. Soc. Chem. Commun., 1244–1245. 25 Sullivan, B.P. and Meyer, T.J. (1986) Organometallics, 5, 1500–1502. 26 Hukkanen, H. and Pakkanen, T. (1986) Inorg. Chim. Acta, 114, L43–L45. 27 Hori, H., Johnson, F.P.A., Koike, K., Ishitani, O., and Ibusuki, T. (1996) J. Photochem. Photobiol. A Chem., 96, 171–174. 28 Koike, K., Hori, H., Ishizuka, M., Westwell, J.R., Takeuchi, K., Ibusuki, T., Enjouji, K., Konno, H., Sakamoto, K., and Ishitani, O. (1997) Organometallics, 16, 5724–5729. 29 Hayashi, Y., Kita, S., Brunschwig, B.S., and Fujita, E. (2003) J. Am. Chem. Soc., 125, 11976–11987. 30 Hori, H., Koike, K., Suzuki, Y., Ishizuka, M., Tanaka, J., Takeuchi, K., and Sasaki, Y. (2002) J. Mol. Catal. A: Chemical, 179, 1–9. 31 Hori, H., Takano, Y., Koike, K., and Sasaki, Y. (2003) Inorg. Chem. Commun., 6, 300–303. 32 Gholamkhass, B., Mametsuka, H., Koike, K., Tanabe, T., Furue, M., and Ishitani, O. (2005) Inorg. Chem., 44, 2326–2336. 33 Tinnemans, A.H.A., Koster, T.P.M., Thewissen, D., and Mackor, A. (1984) Rec. Trav. Chim. Pays-Bas, 103, 288–295.

34 Grant, J.L., Goswami, K., Spreer, L.O., Otvos, J.W., and Calvin, M. (1987) J. Chem. Soc. Dalton Trans., 2105– 2109. 35 Kimura, E., Wada, S., Shionoya, M., and Okazaki, Y. (1994) Inorg. Chem., 33, 770–778. 36 Mochizuki, K., Manaka, S., Takeda, I., and Kondo, T. (1996) Inorg. Chem., 35, 5132–5136. 37 Matsuoka, S., Kohzuki, T., Pac, C., Ishida, A., Takamuku, S., Kusaba, M., Nakashima, N., and Yanagida, S. (1992) J. Phys. Chem., 96, 4437–4442. 38 Matsuoka, S., Yamamoto, K., Ogata, T., Kusaba, M., Nakashima, N., Fujita, E., and Yanagida, S. (1993) J. Am. Chem. Soc., 115, 601–609. 39 Matsuoka, S., Yamamoto, K., Pac, C.J., and Yanagida, S. (1991) Chem. Lett., 2099–2100. 40 Dubois, D.L. (2006) Encyclopedia of Electrochemistry (eds A.J. Bard and M. Stratmann), Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, pp. 202–225. 41 Frese, J.K.W. (1993) Electrochemical and Electrocatalytic Reactions of Carbon Dioxide (eds B.P. Sullivan, K. Krist, and H.E. Guard), Elsevier, Amsterdam, Ch 6. 42 Halmann, M.M., and Steinberg, M. (1999) Greenhouse Gas Carbon Dioxide Mitigation: Science and Technology (eds M.M. Halmann and M. Steinberg), Lewis Publishers, Boca Raton, Ch 12. 43 Hori, Y. (2008) Modern Aspects of Electrochemistry, vol. 42 (eds C.G. Vayenas, R.E. White, and M.E. Gamboa-Aldeco), Springer, New York, pp. 89–189. 44 Taniguchi, I. (1989) Modern Aspects of Electrochemistry, vol. 20 (eds J.M. Bockris, B.E. Conway, and R.E. White), Springer, New York, pp. 327–400. 45 Coehn, A. and Jahn, S. (1904) Chem. Ber., 37, 2836–2842. 46 Rabinowitsch, M. and Maschowetz, A. (1930) Z. Elektrochem., 36, 846–850. 47 Ryu, J., Andersen, T.N., and Eyring, H. (1972) J. Phys. Chem., 76, 3278–3286. 48 Vassiliev, Y.B., Bagotzky, V.S., Osetrova, N.V., Khazova, O.A., and Mayorova, N.A. (1985) J. Electroanal. Chem., 189, 271–294.

313

314

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction 49 Hori, Y., Kikuchi, K., and Suzuki, S. (1985) Chem. Lett., 1695–1698. 50 Koga, O. and Hori, Y. (1993) Electrochim. Acta, 38, 1391–1394. 51 Koga, O. and Hori, Y. (1993) Denki Kagaku, 61, 812–813. 52 Frese, J.K.W. and Canﬁeld, D. (1984) Reduction of CO2 to Methanol and Methane by Photo and Dark Reactions at Semiconductor Electrodes, Gas Research Institute Annual Report. 53 Frese, J.K.W. and Canﬁeld, D. (1985) Reduction of CO2 to Methanol and Methane by Photo and Dark Reactions at Semiconductor Electrodes, Gas Research Institute Annual Report. 54 Frese, K. and Leach, S. (1985) J. Electrochem. Soc., 132, 259–260. 55 Azuma, M., Hashimoto, K., Hiramoto, M., Watanabe, M., and Sakata, T. (1990) J. Electrochem. Soc., 137, 1772–1778. 56 Hoshi, N., Ito, H., Suzuki, T., and Hori, Y. (1995) J. Electroanal. Chem., 395, 309–312. 57 Hoshi, N., Suzuki, T., and Hori, Y. (1996) Electrochim. Acta, 41, 1647–1653. 58 Hoshi, N., Suzuki, T., and Hori, Y. (1997) J. Phys. Chem. B, 101, 8520–8524. 59 Giner, J. (1963) Electrochim. Acta, 8, 857–865. 60 Nikolic, B., Huang, H., Gervasio, D., Lin, A., Fierro, C., Adzic, R., and Yeager, E. (1990) J. Electroanal. Chem., 295, 415–423. 61 Vassiliev, Y.B., Bagotzky, V.S., Khazova, O.A., and Mayorova, N.A. (1985) J. Electroanal. Chem., 189, 295–309. 62 Summers, D., Leach, S., and Frese, K. (1986) J. Electroanal. Chem., 205, 219–232. 63 Seshadri, G., Lin, C., and Bocarsly, A.B. (1994) J. Electroanal. Chem., 372, 145–150. 64 Popic, J., AvramovIvic, M., and Vukovic, N. (1997) J. Electroanal. Chem., 421, 105–110. 65 Bandi, A. and Kuhne, H.M. (1992) J. Electrochem. Soc., 139, 1605–1610. 66 Qu, J., Zhang, X., Wang, Y., and Xie, C. (2005) Electrochim. Acta, 50, 3576–3580.

67 Gattrell, M., Gupta, N., and Co, A. (2006) J. Electroanal. Chem., 594, 1–19. 68 Hori, Y., Kikuchi, K., Murata, A., and Suzuki, S. (1986) Chem. Lett., 897–898. 69 Ikeda, S., Amakusa, S., Noda, H., Saito, Y., and Ito, K. (1988) Proceedings of the Electrochemical Society, section: Photoelectrochemical and Electrosynthesis on Semiconducting Materials, Vol. 88 (14), pp. 130–136. 70 Noda, H., Ikeda, S., Oda, Y., and Ito, K. (1989) Chem. Lett., 289–292. 71 Cook, R.L., Macduff, R.C., and Sammells, A.F. (1987) J. Electrochem. Soc., 134, 1873–1874. 72 Dewulf, D.W., Jin, T., and Bard, A.J. (1989) J. Electrochem. Soc., 136, 1686–1691. 73 Kim, J.J., Summers, D.P., and Frese, K.W. (1988) J. Electroanal. Chem., 245, 223–244. 74 Oda, I., Ogasawara, H., and Ito, M. (1996) Langmuir, 12, 1094–1097. 75 Nogami, G., Itagaki, H., and Shiratsuchi, R. (1994) J. Electrochem. Soc., 141, 1138–1142. 76 Shiratsuchi, R., Aikoh, Y., and Nogami, G. (1993) J. Electrochem. Soc., 140, 3479–3482. 77 Jermann, B. and Augustynski, J. (1994) Electrochim. Acta, 39, 1891–1896. 78 Eggins, B.R. and McNeill, J. (1983) J. Electroanal. Chem., 148, 17–24. 79 Isse, A.A., Gennaro, A., Severin, M.G., and Vianello, E. (1993) Proceedings, International Conference on Carbon Dioxide Utilization, Bari, Italy, pp. 287–294. 80 Ikeda, S., Takagi, T., and Ito, K. (1987) Bull. Chem. Soc. Jpn., 60, 2517–2522. 81 Ito, K., Ikeda, S., Yamauchi, N., Iida, T., and Takagi, T. (1985) Bull. Chem. Soc. Jpn, 58, 3027–3028. 82 Kaiser, U. and Heitz, E. (1973) Ber. Bunsenges. Phys. Chem., 77, 818–823. 83 Desilvestro, J. and Pons, S. (1989) J. Electroanal. Chem., 267, 207–220. 84 Aylmer-Kelly, A.W.B., Bewick, A., Cantrill, P.R., and Tuxford, A.M. (1974) Faraday Discuss. Chem. Soc., 56, 96–105. 85 Haynes, L.V. and Sawyer, D.T. (1967) Anal. Chem., 39, 332–338.

References 86 Benson, E.E., Kubiak, C.P., Sathrum, A.J., and Smieja, J.M. (2009) Chem. Soc. Rev., 38, 89–99. 87 Hawecker, J., Lehn, J.M., and Ziessel, R. (1984) J. Chem. Soc. Chem. Commun., 328–330. 88 Sullivan, B.P., Bolinger, C.M., Conrad, D., Vining, W.J., and Meyer, T.J. (1985) J. Chem. Soc. Chem. Commun., 1414–1415. 89 Ishida, H., Tanaka, H., Tanaka, K., and Tanaka, T. (1987) J. Chem. Soc. Chem. Commun., 131–132. 90 Ishida, H., Tanaka, K., and Tanaka, T. (1985) Chem. Lett., 405–406. 91 Ishida, H., Tanaka, K., and Tanaka, T. (1987) Organometallics, 6, 181–186. 92 Bruce, M.R.M., Megehee, E., Sullivan, B.P., Thorp, H., Otoole, T.R., Downard, A., and Meyer, T.J. (1988) Organometallics, 7, 238–240. 93 Bruce, M.R.M., Megehee, E., Sullivan, B.P., Thorp, H.H., Otoole, T.R., Downard, A., Pugh, J.R., and Meyer, T.J. (1992) Inorg. Chem., 31, 4864– 4873. 94 Bolinger, C.M., Story, N., Sullivan, B.P., and Meyer, T.J. (1988) Inorg. Chem., 27, 4582–4587. 95 Dubois, D.L., Miedaner, A., and Haltiwanger, R.C. (1991) J. Am. Chem. Soc., 113, 8753–8764. 96 Bernatis, P.R., Miedaner, A., Haltiwanger, R.C., and Dubois, D.L. (1994) Organometallics, 13, 4835–4843. 97 Fisher, B. and Eisenberg, R. (1980) J. Am. Chem. Soc., 102, 7361–7363. 98 Beley, M., Collin, J.P., Ruppert, R., and Sauvage, J.P. (1984) J. Chem. Soc. Chem. Commun., 1315–1316. 99 Beley, M., Collin, J.P., Ruppert, R., and Sauvage, J.P. (1986) J. Am. Chem. Soc., 108, 7461–7467. 100 Hay, R.W., Crayson, J.A., Cromie, T.J., and Lightfoot, P. (1997) Polyhedron, 16, 3557–3563. 101 Balazs, G.B. and Anson, F.C. (1992) J. Electroanal. Chem., 322, 325–345. 102 Rudolph, M., Dautz, S., and Jager, E. (2000) J. Am. Chem. Soc., 122, 10821–10830. 103 Reda, T., Plugge, C.M., Abram, N.J., and Hirst, J. (2008) Proc. Natl Acad. Sci. USA, 105, 10654–10658.

104 Halmann, M. (1978) Nature, 275, 115–116. 105 Halmann, M.M. and Steinberg, M. (1999) Greenhouse Gas Carbon Dioxide Mitigation: Science and Technology (eds M.M. Halmann and M. Steinberg), Lewis Publishers, Boca Raton, Ch 13. 106 Inoue, T., Fujishima, A., Konishi, S., and Honda, K. (1979) Nature, 277, 637–638. 107 Taniguchi, I., Aurianblajeni, B., and Bockris, J.O. (1983) J. Electroanal. Chem., 157, 179–182. 108 Taniguchi, I., Aurianblajeni, B., and Bockris, J.O. (1984) J. Electroanal. Chem., 161, 385–388. 109 Yoneyama, H., Sugimura, K., and Kuwabata, S. (1988) J. Electroanal. Chem., 249, 143–153. 110 Canﬁeld, D. and Frese, J.K.W. (1983) J. Electrochem. Soc., 130, 1772–1773. 111 Frese, J.K.W. and Canﬁeld, D. (1984) J. Electrochem. Soc., 737, 2518–2522. 112 Ikeda, S., Yoshida, M., and Ito, K. (1985) Bull. Chem. Soc. Jpn, 58, 1353–1357. 113 Ikeda, S., Saito, Y., Yoshida, M., Noda, H., Maeda, M., and Ito, K. (1989) J. Electroanal. Chem., 260, 335–345. 114 Hinogami, R., Nakamura, Y., Yae, S., and Nakato, Y. (1998) J. Phys. Chem. B, 102, 974–980. 115 Kaneco, S., Katsumata, H., Suzuki, T., and Ohta, K. (2006) Appl. Catal. B: Environmental., 64, 139–145. 116 Aurianblajeni, B., Taniguchi, I., and Bockris, J.O. (1983) J. Electroanal. Chem., 149, 291–293. 117 Cabrera, C.R. and Abruna, H.D. (1986) J. Electroanal. Chem., 209, 101–107. 118 Daube, K.A., Harrison, D.J., Mallouk, T.E., Ricco, A.J., Chao, S., Wrighton, M.S., Hendrickson, W.A., and Drube, A.J. (1985) J. Photochem., 29, 71–88. 119 Taniguchi, Y., Yoneyama, H., and Tamura, H. (1982) Bull. Chem. Soc. Jpn, 55, 2034–2039. 120 Bockris, J.O. and Wass, J.C. (1989) J. Electrochem. Soc., 136, 2521–2528. 121 Beley, M., Collin, J.P., Sauvage, J.P., Petit, J.P., and Chartier, P. (1986) J. Electroanal. Chem., 206, 333–339. 122 Petit, J.P., Chartier, P., Beley, M., and Deville, J.P. (1989) J. Electroanal. Chem., 269, 267–281.

315

316

11 Photochemical, Electrochemical, and Photoelectrochemical Reduction 123 Petit, J.P., Chartier, P., Beley, M., and Sauvage, J.P. (1987) New J. Chem., 11, 751–752. 124 Parkinson, B.A. and Weaver, P.F. (1984) Nature, 309, 148–149. 125 Barton, E.E., Rampulla, D.M., and Bocarsly, A.B. (2008) J. Am. Chem. Soc., 130, 6342–6344. 126 Rossetti, R., Nakahara, S., and Brus, L.E. (1983) J. Chem. Phys., 79, 1086–1088. 127 Yamamura, S., Kojima, H., Iyoda, J., and Kawai, W. (1987) J. Electroanal. Chem., 225, 287–290. 128 Henglein, A., Gutierrez, M., and Fischer, C.H. (1984) Ber. Bersenges. Phys. Chem., 88, 170–175. 129 Eggins, B.R., Robertson, P.K.J., Stewart, J.H., and Woods, E. (1993) J. Chem. Soc. Chem. Commun., 349–350. 130 Kuwabata, S., Nishida, K., Tsuda, R., Inoue, H., and Yoneyama, H. (1994) J. Electrochem. Soc., 141, 1498–1503.

131 Eggins, B.R., Irvine, J.T.S., Murphy, E.P., and Grimshaw, J. (1988) J. Chem. Soc. Chem. Commun., 1123–1124. 132 Nedeljkovic, J.M., Nenadovic, M.T., Micic, O.I., and Nozik, A.J. (1986) J. Phys. Chem., 90, 12–13. 133 Yamamura, S., Kojima, H., Iyoda, J., and Kawai, W. (1988) J. Electroanal. Chem., 247, 333–337. 134 Cook, R.L., Macduff, R.C., and Sammells, A.F. (1988) J. Electrochem. Soc., 135, 3069–3070. 135 Halmann, M.M. and Steinberg, M. (1999) Greenhouse Gas Carbon Dioxide Mitigation: Science and Technology (eds M.M. Halmann and M. Steinberg), Lewis Publishers, Boca Raton, Ch 14. 136 Adachi, K., Ohta, K., and Mizuno, T. (1994) Sol. Energy, 53, 187–190.

317

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation Based on Carbon Dioxide Giuseppe Silvestri and Onofrio Scialdone

12.1 Introduction

The use of electrochemistry for synthetic or structural studies involving carbon dioxide (CO2) dates back to the very start of the nineteenth century, following Volta’s invention of the pile [1]. Some seventy years later, the ﬁrst reproducible results of the direct electrochemical reduction of CO2 in aqueous media were reported, almost contemporarily, by Beketov [2] and Royer [3]. The use of sodium or potassium, whether pure or as amalgams, for the transformation of CO2 in the presence of water to formic acid, was reported at almost the same time [4]. Since then, a major research effort – initially with the aim of reproducing (using simple laboratory methods) the natural process of growth of plants – has been addressed towards the reconversion of CO2 to organic compounds. During the second half of the twentieth century, the electroreduction of CO2 to chemicals was driven not only by a wide variety of scientiﬁc and technological advances, but also by an increasing public opinion regarding the accumulation of “greenhouse gases” in the Earth’s upper atmosphere. The extraordinary technical progress which has been made in the conversion of solar to electric energy, associated with increased selectivities in the electroreduction of CO2 in aqueous media, has led to the hypothesis of a “green route” for producing chemicals that can be either stored or reused for energetic purposes [5]. At almost the same time, a second area of research, based on nonaqueous electrochemical processes, has given rise to some interesting synthetic data. Indeed, the latter approach, in which CO2 is used as a building block, has proved to be successful for the synthesis of a variety of classes of carboxylated derivatives. The details of electrosynthetic processes in nonaqueous solvents were ﬁrst reviewed almost two decades ago [6, 7], and a general survey of this rather fragmented ﬁeld might be useful when proposing this methodology as a valuable synthetic tool. However, the situation changes when considering the much wider ﬁeld of electroreductions in protic, mainly aqueous, media. As this area has been reviewed extensively in the past, attention at this point will be limited to the latest contributions, with frequent reference to the studies of Gattrel and Gupta [8]. Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

318

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

12.2 Electrocarboxylation

Carbon dioxide represents a central and cheap building block for organic synthesis. Its facile reaction with carbanions or procarbanions, as in Grignard reactions, allows the straightforward attachment of CO2 to an organic skeleton in order to increase it by one carbon unit, while maintaining an open functionalization. The electrochemical reduction of several types of organic compound – such as halides or pseudohalides, ketones and aldehydes, alkenes, and alkynes – represents a convenient and cheap method of generating carbanions or procarbanions, in stoichiometric fashion. In fact, the electrochemical route presents a major advantage of not having to use phosgene and cyanides, and this is perhaps the main reason why much effort has been expended during the past ﬁfty years in studying the mechanistic and preparative aspects of the electrocarboxylation of several substrates involving alkenes, alkynes, polyenes, ketones, aldehydes, imines, halides, and thioethers. During the last two decades of the twentieth century, several extended reviews have been produced detailing a large amount of information on this subject [6, 7]. Likewise, many private companies have investigated in detail the application potential of the electrocarboxylation of many materials (mainly aromatic ketones and benzyl halides), and this has led to the creation of pilot plants for industrial application. A particular case has been in the production of antiinﬂammatory drugs. Despite so many studies having been conducted, there appear to be no industrial processes that currently involve electrocarboxylation reactions. Yet, it cannot be excluded that such industrial processes might exist, as in many cases such productions (e.g., of ﬁne chemicals) may have gone unrecognized for reasons of industrial secrecy [9]. Clearly, during the past few years a strong interest has been maintained in the subject of electrocarboxylation, as attested by the vast number of reports [10–80] and patents [81] that have been prepared, from both academic institutions and private companies. In general, electrocarboxylation reactions are carried out in aprotic solvents such as acetonitrile (ACN), N,N-dimethylformamide (DMF) or N-methyl-2-pyrrolidone (NMP) in a one-compartment cell by the use of sacriﬁcial anodes (Al or Mg), as the use of these systems generally provide important advantages [7, 10–12] that include:

•

Mono- or poly-carboxylated anions always gives rise to complex salts with the metal cations, which are insoluble in the reaction medium, or can be made insoluble by addition of appropriate cosolvents, making the isolation procedure more easy.

•

The separator is eliminated, thus avoiding various serious problems related to the use of membranes, such as a the high ohmic resistance.

•

For some substrates (e.g., benzyl halides) the metal cations can complex some of the reaction intermediates or ﬁnal products, thus avoiding or minimizing undesirable side reactions.

12.2 Electrocarboxylation

•

The operation of the cell in a single-compartment mode with a sacriﬁcial anode provokes the electrogeneration of ions in proportion to the charge passed. These can sustain the current ﬂow in the cell, so that there is usually no need to operate in the presence of a high concentration of added supporting electrolyte in batch cells.

Recently, the role of metal cations has been studied in detail [13, 14], it having been observed that when Al or Mg cations are added to the reaction medium, they have a strong inﬂuence on the reduction process of alkynes, ketones and halides, and in most cases this leads to a complete change of the mechanism. In contrast, when ions are supplied to the reaction medium by oxidation of the anode during electrocarboxylation, they show no effect on the cathodic process. Hence, it appears that metal cations produced by the consumable anode in undivided cells are intercepted and complexed by the carboxylated products before they can reach the cathode. Therefore, the cathodic process can occur in a medium without free metal ions. In the following sections, attention is focused on recent results relating to the electrocarboxylation of the most frequently investigated substrates, notably aromatic ketones and benzyl halides. A brief summary of some studies involving other substrates is also provided. The electrocarboxylation of aromatic ketones and benzyl halides has attracted much interest for two main reasons. First, the production of alpha-aryl-propionic acids used as nonsteroidal anti-inﬂammatory drugs (NSAIDs), can be achieved from the corresponding benzyl halides, or from the ketones [7]. Second, the electrocarboxylation of these substrates, in bench-scale systems, has provided promising results that have prompted an in-depth investigation of these synthetic routes. 12.2.1 Electrocarboxylation of Organic Halides

The electrocarboxylation of organic halides ideally involves the following reaction scheme [21]: RX + e− → R • + X −

(12.1)

RX + e − RX •−

(12.2a)

RX → R + X

(12.2b)

•−

•

−

R• + e− R −

(12.3)

R + CO2 → RCO −

− 2

(12.4)

in which RX is the organic halide or pseudohalide. The electrocarboxylation process involves a reductive cleavage of the carbon–halogen bond, followed by a further reduction of the generated radical to the corresponding carbanion, which is then captured by CO2 (Equations 12.1–12.4). There are two possible mechanisms for the dissociative electron transfer to RX: (i) a concerted mechanism in which

319

320

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

electron transfer (ET) and bond breaking occur in a single step (Equation 12.1); or (ii) a stepwise mechanism (12.2) in which a labile radical anion RX• − is initially formed [15]. In the case of benzyl chlorides, such as that investigated in most reported studies, the bond breaking is expected to occur in a single step [16, 17]. Quite often, R• is more easily reduced than RX, so that it is reduced to R− at the working potential as soon as it is formed. It is important to note that in most of these reported studies the electrocarboxylation of organic halides occurs in an undivided cell by the use of a sacriﬁcial anode (Al, Mg). Indeed, the small cations metn+ generated at the anode may readily complex the formed carboxylate: met (anode ) − ne − met n + met

n+

(12.5)

+ n RCO2− met (RCO2 )n

(12.6)

so that any possible esteriﬁcation of the parent RX (Equation 12.7) is prevented, thus allowing the achievement of higher yields of carboxylated products [10–13]: RX + RCO2-

RCOOR + X-

(12.7) −

Presumably, this is also true for the coordination of R , so that this strong base is kinetically stabilized and can wait to react almost exclusively with CO2 instead of being partially protonated in the medium. During the past twenty years, the electrocarboxylation of numerous organic halides has been carried out in undivided cells by direct reduction at various cathodes, including generally carbon, platinum, or Hg cathodes [7, 16, 18–20, 81]. One drawback of this method of electrosynthesis is that reduction of the organic halide may require very negative potentials. Indeed, this is the case with most chlorides, which are more interesting from an applications point of view, or, in general, when the carbon–halogen bond is not activated by the presence of electron-withdrawing groups or by the proximity of a π* system [21]. In such cases, the process may involve the concurrent reduction of CO2 (Equations 12.8 and 12.9), which yields several byproducts, thus lowering the current efﬁciency and creating some separation problems. CO2 + e − CO•− 2

(12.8)

CO → products

(12.9)

•− 2

As noted above, the electrocarboxylation process has been widely investigated for the production of some NSAIDs, starting from parent arylethyl halides [6, 7]. On the other hand, the direct process at conventional cathodes requires quite negative potentials, and gives rise in some cases to moderate faradic efﬁciencies. Furthermore, attempts to scale-up the process in the case of the reduction of 1-(3-phenoxyphenyl)-1-chloroethane [18] and 1-(4-isobutylphenyl)-1-chloroethane [16] (which are the precursors of fenopren and ibuprofen, respectively) gave very different results with respect to those obtained in syntheses performed in benchscale systems. In particular, passivation of the cathode surface was observed, and this resulted in lower yields and selectivities. Similar results were also observed during the electrocarboxylation of chloroacetonitrile [19].

12.2 Electrocarboxylation

Two homogeneous catalytic systems have been proposed to allow electrocarboxylation at a much less negative electrode potential than is required for RX reduction, and far from the cathodic surface. The ﬁrst consists of using a transition metal catalyst such as Ni, Pd and Co [6, 7, 22–27]. Various nickel [22–24] and palladium [26] complexes with phosphine ligands have demonstrated good catalytic properties in the electrocarboxylation of aromatic and benzylic halides. For example, in the case of Ni or Pd, the reduction of MeII generates a zerovalent metal center Me0 which is added oxidatively to the organic halide or pseudohalide to generate an organometal, RMeIIX. This species reduces generally at potentials much less negative than that of the parent organic halide, due to the activation by the metal(II) center, giving rise to an organometal, RMeI (Equation 12.12) or an organometal(0) anion (Equation 12.13), respectively, depending on the nature of the metal catalyst and of its ligands: Me II + 2 e − → Me 0

(12.10)

Me 0 + RX → RMe II X

(12.11)

RMe X + e → RMe + X II

−

I

−

(12.12)

RMe II X + 2 e − → RMe 0 − + X −

(12.13)

In the ﬁrst case, for example, with NiCl2(dppe) as the catalyst, CO2 oxidatively adds to the organometal(I) center to yield R–MeIII(CO2), which gives rise to the formation of a carboxylate anion and of a metal(I) species that is reduced back to Me0 so that a new cycle can start (Equations 12.14–12.16): RMe I + CO2 → RMe III(CO2 )

(12.14)

RMe (CO2 ) → RCO + Me

(12.15)

III

− 2

I

Me I + e − → Me 0

(12.16) 0

In the second case, for example, with PdCl2(PPh3)2 as the catalyst, RMe gives rise to MeII which is reduced to the zerovalent metal species (Equation 12.10) and to the anion R− (Equation 12.17) which reacts with CO2 (Equation 12.4) [21]. RMe 0 − R − + Me II

( ΔG° >> 0 )

(12.17)

Both routes have been studied extensively, and their mechanism widely investigated [22–26]. Square–planar cobalt complexes such as CoII(salen) (H2salen = N,N′-bis(salicylidene)-ethane-1,2-diamine) [27] and Co(tetraphenylporphyrin) [28] have displayed catalytic activity in the electrocarboxylation of several alkyl halides and benzyl halides. The reaction between electrogenerated CoI and RX yields an organometallic intermediate, [CoIII(L)(R)] which, upon further reduction, releases the alkyl group as a radical. These synthetic paths generally are very selective as the metal center participation avoids or minimizes the formation of byproducts. Indeed, the radical R. and anion R− are not produced, or R− is “protected” under the RMe0− form, due to the large endergonicity of equilibrium (Equation 12.17), so that it can react only with the strongest electrophiles (viz. CO2 here). In spite

321

322

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

of the good yields achieved with several metal complexes, this method has the major disadvantage of requiring expensive catalytic materials, and in some cases also poorly conducting solvents or unusual electrolytes (which imposes large ohmic drop losses in a large cell) [21]. The second alternative to bypass a difﬁcult RX reduction consists of using redox catalysis [29]. Thus, the reduction of RX can be performed at the much less negative one-electron reversible reduction potential (Equation 12.18) of an adequate redox mediator M, which delivers the electron to RX through an homogeneous electron transfer when RX.− does exist (Equation 12.19), or for a concerted bondbreaking RX reduction (Equation 12.20): M + e− → M•−

(12.18)

M•− + RX M + RX •− (→ R • + X − )

(12.19)

M + RX → M + R + X

(12.20)

•−

•

−

M•− + R • M + R − •

(12.21) −

The radical R is, then, converted into R (Equation 12.21) which reacts with CO2 to yield the organic carboxylate (Equation 12.4). Hence, the target product is electrogenerated at the reduction potential of the redox mediator catalyst. Interestingly, in this process since the electrode processes involve only the reduction of the extremely chemically stable M/M• − redox couple, the cathode deactivation should be prevented. Indeed, it has been shown, in the case of benzyl chlorides, that it is possible to hinder the electrode passivation and to increase ﬁnal yields by using suitable homogeneous charge-transfer catalysts (HCTCs) [16]. These results have prompt a more extensive research on the use of HCTCs for the electrocarboxylation of benzyl halides [21, 30]. In particular, it has been shown that a very carefully selection of the catalyst and of operative conditions must be performed in order to avoid or minimize some important potential disadvantages of this methodology, such as the possible homogeneous reduction of CO2 by means of HCTCs (Equation 12.22) and a rapid decomposition of the catalyst [21]. M•− + CO2 M + CO•− 2

(12.22)

Thus, in the catalyzed process some slow catalyst decomposition reactions, although unnoticeable at the voltammetric level, can lead to a relevant deactivation of the catalyst at a preparative-scale level. Interestingly, the catalyst deactivation can be, counterintuitively, minimized by operating at sufﬁciently high [RX]bulk/[catalyst] ratio. The competition between CO2 and substrate homogeneous reduction has been shown to depend drastically on the redox standard potential of the mediator, thus allowing the minimization of the reaction reported in equation 12.21 by operating with a catalyst that reduces at a sufﬁciently positive redox potential [16]. A third catalytic route consists of using cathode materials with catalytic properties. The electrocatalytic activity of electrode materials towards the reduction of organic halides has been the object of many studies during the past few years [82], with silver having been shown to possess powerful electrocatalytic activities

12.2 Electrocarboxylation

towards the reduction of a large variety of organic halides, including benzyl and benzylic-type chlorides. In particular, it has been shown recently that it is possible to perform the electrocarboxylation of benzyl and benzylic-type chlorides [31–33], halopyridine [34], chloroacetonitrile [35] and haliphatic halides [36] by direct reduction on silver cathodes at potentials that are generally dramatically less negative than that recorded at carbon and mercury electrodes, and similar to those used in homogeneous catalytic systems. Recently, it has also been shown that a scale-up of the electrocarboxylation of benzyl chlorides at silver gave similar results to those of bench-scale experiments, without any passivation or disaggregation of the cathode [33]. A comparison between the performances of electrocarboxylation processes of benzyl halides at silver anodes and by the use of HCTCs, has been carried out by performing the electrolysis of 1-phenyl-1-chloroethane in NMP in the presence of aluminum as a sacriﬁcial anode [37]. The data listed in Table 12.1 show that, when the process is performed with HCTCs, higher selectivity and faradic efﬁciency of the target product are obtained. In contrast, in the catalyzed process the synthetic system is complicated by the presence of the catalyst as an additional component that is, furthermore, partially consumed during the electrolysis. Simple kinetic models for the direct and homogeneous electrocarboxylation of benzyl halides were recently developed based on competitions at two successive stages involving R• and R−, respectively, and on the competition between reduction processes involving CO2 and RX. Interestingly, the experimental data obtained were in satisfactory agreement with theoretical predictions, both in the case of the direct process [33] (in spite of the fact that no adjustable parameters were used in the model) and of the process mediated by a HCTC [21]. Recently, the electrocarboxylations of benzyl and aryl halides and perﬂuoroalkylhalides [39] in supercritical mixture or in supercritical carbon dioxide (scCO2) and of aryl and benzyl halides in microemulsion [40], were also investigated in order to exploit the possible effect of the use of these solvents on the selectivity of the

Table 12.1

Electrocarboxylation of 0.07 M 1-phenyl-1-chloroethane in NMP. Carbon

Silver

HCTC

Working potential (vs. SCE)

−2.35

−1.6–1.8

−1.95

Initial substrate and catalyst concentrations

[RX] °:0.05–0.15 M

[RX] °:0.14–0.28 M

[RX]: 0.18–0.22 M [M] °: 5 mM

Conversion, Faradaic efﬁciency (%) Selectivity (%)

<50 ∼50–60 ∼72–76

>90 ∼76–82 ∼80–83

>90 ∼90–92 ∼90–95

Catalyst consumption (M °–M)/M °

20%

323

324

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

process. The possibility of inducing, by electrochemical means, the diastereoselective carboxylation of halogenated organic compounds was investigated, and provided interesting results [38]. The studies included the diastereoselective electrochemical carboxylation of chiral N-(2-bromoacyl)oxazolidin-2-ones to chiral alkylmalonic acid derivatives, which are used as building blocks in the synthesis of molecules with biological activity, and of chiral propane-1,3-diols derivatives [38a]. The reaction was carried out via a cathodic reduction of the C–Br bond, in the presence of CO2, followed by treatment with diazomethane. Under the best conditions, a good yield of the carboxylated product was obtained (80%) with excellent diastereoselectivity (98 : 2). Others have investigated the inﬂuence of the presence of β-cyclodextrin in the reaction medium on the electrochemical carboxylation of α-bromoethylbenzene and 1-(4-isobutylphenyl)ethyl chloride [41]. It has been reported that the preparative electrocarboxylation of the inclusion complex β-cyclodextrin-1-(4-isobutylphenyl) ethylchloride afforded the S-form of 2-(4-isobutylphenyl) propionic acid (S-ibuprofen) in a high enantiomeric excess (97%). 12.2.2 Electrocarboxylation of Aromatic Ketones

The electrocarboxylation of aldehydes and ketones leads to the corresponding αhydroxycarboxylic acids that can easily be converted into carboxylic acids via a hydrogenation reaction [7]. It has been reported that the electrocarboxylation of aromatic ketones occurs through the reaction of CO2 onto the activated carbon atom of the carbonyl group of the ketyl radical anion generated upon electron transfer to the ketone [7]. Otherwise, the aforementioned intermediate is likely to be a resonance hybrid (see Equation 12.23), and its electrophilic reaction with CO2 may take place both at the carbon or the oxygen atom [42, 43]. RR ′C = O + e − = RR ′C•-O− (Ia ) ↔ RR ′C− -O•(Ib )

(12.23)

Hydroxcarboxylic acids of interest may be formed through these two different reaction pathways, all of which involve the consumption of two electrons per molecule of substrate carboxylated (see for example Equations 12.24–12.26).

[R1R 2CO]•− + CO2 → [R1R 2CO-COO]•− •−

(12.24) 2−

[R1R 2CO-COO] + e − → [R1R 2CO-COO] 2−

[R1R 2CO-COO] + CO2 → OOCO- (R1R 2 ) C-COO −

(12.25) −

(12.26)

Generally, it seems reasonable to assume that both resonance forms can play a role (Equation 12.23), and that the charge density on Ia increases in the presence of electron-donating groups. On the other hand, when considering side reactions, the predominance of form Ia or Ib is likely to have a dramatic inﬂuence on the selectivity of the process. For example, the presence of a high spin density on the carbon atom may speed-up any radical–radical coupling reaction, thereby increasing the formation of unwanted pinacolic dimers [43].

12.2 Electrocarboxylation

An electrocarboxylation methodology based on the use of un undivided cell, equipped with sacriﬁcial anodes, was successfully used for the synthesis of 2-hydroxy-2-(6-methoxy-2-naphthyl)propionic acid and 2-hydroxy-2-(4-i-butylphenyl)propionic acid, the precursors of naproxen and ibuprofen, respectively [7]. Interestingly, the production costs of naproxen, on the basis of the abovementioned technology, were found to be competitive with the existing procedures, despite the fact that the electrochemical-based route did not use phosgene or cyanides [44]. The isolation procedure was complicated by the fact that the aluminum salts were soluble in the reaction medium. In order to obtain an almost quantitative separation of the products, precipitation with alcohol or ethers can be used [10, 44, 45]. Others [45, 46] have investigated the inﬂuence of various operative parameters on the process, and found it possible to achieve good yields in the target product at pilot scale. In recent years, several investigators have shown that the process of electrocarboxylation of aromatic ketones, in terms of selectivity and faradic efﬁciency in the target product, depend dramatically on the nature of the substrate and on the adopted operative conditions [46–49, 51, 52]. As an example, lower yields in the carboxylate are generally obtained with substituted alkyl phenyl ketones, whereas high yields with the corresponding benzophenones and intermediate result in 6-methoxy-2-acetonaphthone [46, 47]. For the above-mentioned substrates, under the adopted experimental conditions, the most relevant products of the synthesis were the hydroxy acids and the corresponding alcohols and pinacols. Otherwise, the ketyl radical anion of aromatic ketones bearing reactive groups (e.g., as in the case of halogenated compounds) can give rise also to other side reactions, such as cleavage of the carbon–halogen bond [49, 50]. In particular, in the case of halogenated aromatic ketones, a higher selectivity of the target products was obtained with chlorobenzophenones with respect to the homo-substituted chloroacetophenones as a result of a less relevant formation of dehalogenated compounds, alcohols and pinacols, by changing from the ortho to para to meta isomer, and from the bromo to chloro to ﬂuoro derivative [49b]. It has also been shown recently that an unexpected ring carboxylation could occur during the electrocarboxylation of aromatic ketones in anhydrous media [47]. Fortunately, low yields of these products have been observed for most of the investigated ketones. On the other hand, these side processes may lead to increased costs for the ﬁnal separation of pure carboxylic acids. Recently, speciﬁc studies were devoted to determining a set of operative conditions that would allow an optimization of selectivity and faradic efﬁciency of the target hydroxy acids [51, 52]. In particular, a high ratio between the concentrations of CO2 and the ketone, achieved for example by using a moderately high CO2 pressure, favored the formation of carboxylated products with respect to side reactions [51], including ring carboxylation processes [52], that could be minimized also by adding small amounts of water to the reaction medium. Recently, the effect of the cathode material on the target product yield was also studied in detail [53, 81a], by using also diamond ﬁlm cathode in a divided cell [81a]. An attempt to produce atrolactic acid via an asymmetric electrochemical carboxylation of prochiral acetophenone was also recently reported [54].

325

326

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

12.2.3 Electrocarboxylation of Other Substrates

As noted above, during the past few years the electrocarboylation of several other substrates [6–7, 55–79] involving alkenes [55–58], alkynes [58–60], aldehydes [61], aliphatic ketones [62], imines [63], aryldiazonium tetraﬂuoroborates [63], derivatives of cinnammic acid [64–66] has been investigated by many groups. Dunach, Inesi and others also investigated the electrochemical synthesis of cyclic carbonates from CO2 with epoxides, alcohols and glycols [66]. In this regard, Yang et al. [67] reported the use of pure room temperature ionic liquids (ILs) as reaction media in the electrochemical activation of CO2 for the synthesis of cyclic carbonate from epoxide, under mild conditions. CO2-saturated IL (BMIMBF4) solutions were also used for the electrochemical carboxylation of activated oleﬁns [68]. Monocarboxylic acids were obtained in moderate yield (35–55%), and the IL was recycled ﬁve times. Wang et al. [64] investigated the effect of various operative parameters on the electrochemical synthesis of cyano-phenyl-propionic acid by carboxylation of cinammonitrile in undivided cells equipped with Mg anodes and a Ni cathode. Inesi et al. [65] investigated the possibility of synthesizing chiral 2-phenyl succinic ester derivatives by the electrochemical reduction of chiral cinnamic acid derivatives under a CO2 atmosphere. When 4R-(diphenylmethyl)-oxazolidin-2-one was used as a chiral auxiliary, the two diastereoisomers could be easily separated by using ﬂash chromatography, and the R-isomer was obtained as the major product. Attention has also been paid during the past few years to examining the electrochemical synthesis of cyanoacetic acid, a product which is of relevant interest to many industrial processes, and is currently produced via an industrial scheme based on the reaction between chloroacetic acid and dangerous reagents such as alkali metal cyanides. The electrocarboxylation of chloroacetonitrile at Hg [69], carbon [19, 69] or silver [35] cathodes gives rise to quite high current efﬁciencies in the target product when suitable operative conditions are adopted; however, the process suffers from the inevitable use of highly negative working potentials at the Hg and carbon cathodes. An alternative electrochemical methodology for the synthesis of cyanoacetic acid was based on the electrolysis of a solution containing acetonitrile, R4NX salts, and CO2. Although this method was very simple, it required very negative potentials (<−2.3 V versus SCE) and gave rise to quite low faradaic efﬁciencies [70]. The best results, obtained with an experimental set-up using undivided cells equipped with sacriﬁcial anodes, showed faradic yields of up to 30–40% [70c]. Barba et al. [71] also studied the synthesis of α-methyl and α-ethylcyanoacetic acids from propionitrile and butyronitrile, respectively, through a paired electrochemical reaction with CO2. Inesi and coauthors showed that a carboxylating reagent could be obtained by bubbling CO2 into a CH3CN-TEAP (tetraethylammonium perchlorate) solution that had previously been electrolyzed [72]. Subsequently, organic carbamates were isolated in high yields from these solutions following the addition of amines and an alkylating agent [72a].

12.3 The Electroreduction of Carbon Dioxide in Protic Media (Water and Alcohols)

Numerous studies have been dedicated to the electrocarboxylation of alkenes and alkines. For example, in 2001 Dunach reported the electrochemical carboxylation of terminal alkynes and diynes on silver cathodes in one-compartment cells ﬁtted with magnesium anodes to afford selectively monocarboxylic acid derivatives in good yields [60]. The group of Tokuda has also investigated the electrocarboxylation of many substrates [73–77], including allylic [73], vinylic [74], and propargylic bromides. Tokuda pointed out that the electrochemical carboxylation of allylic bromides would allow the regioselective synthesis of β,γ-unsaturated acids by using a platinum cathode and a magnesium anode [73]. When, in the vinylic molecule (R1R2CCR3Br), R1, R2 or R3 are alkyl or hydrogen atoms, the reduction potentials of the vinyl bromides become highly negative and their electrochemical carboxylations give rise to α,β-unsaturated carboxylic acids with low yields [74a] that can be enhanced by the addition of a Ni(II) catalyst [74b]. The electrochemical carboxylation of (E)- and (Z)-β-bromostyrene in the presence of 20 mol% NiBr (bpy) proceeded with a retention of the stereochemistry to yield the corresponding (E)- and (Z)-cinnamic acids with high stereoselectivities [74c]. Tokuda and coworkers studied also the electrocarboxylation of vinyl triﬂates [76]; in these molecules, the group –OSO2CH3 acted as a good leaving group. In the electrochemical carboxylation of vinyl triﬂates two different products, α.β-unsaturated carboxylic acids or β-keto carboxylic acids were generally obtained using as substrates the phenyl-substituted and alkyl-substituted vinyl triﬂates, respectively. Interestingly, these authors also studied the electrocarboxylation of various substrates in a supercritical mixture of CO2 and acetonitrile [39a, b, 77], while Grimberg et al. [78] reported the electrocarboxylation of 1,4-dibromobut-2-ene in a CO2–DMF liquid mixture.

12.3 The Electroreduction of Carbon Dioxide in Protic Media (Water and Alcohols)

The stepwise electron reduction of CO2, whether direct or indirect, catalyzed, or by direct transfer on an apparently inert conductive surface, has been the object of considerable attention since the ﬁrst concise reports of formate anion production. Since then, the list of possible derivatives has grown from formates to carbon monoxide, methane, ethylene, and short-chain saturated hydrocarbons. As noted in Section 12.1, this area of research has been expanded in recent years [8, 80, 83], with information relating to increased yields, to the effect of electrode materials on selectivity, as well as further speculations on possible reaction mechanisms, having been obtained on a continuous basis. Yet, the key to these synthetic processes – an understanding of the relationship between the surface of the electrode and the synthetic behavior of the system – seems no closer to being identiﬁed. The following brief review of recent reports of this subject will focus on the nature of the cathode surface, for which some surprising results have been attained. Attention will also be focused on the hypotheses of applications of the various proposed systems, rather than on the mechanisms or on further electroanalytic

327

328

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

investigations. Copper, with its moderate hydrogen overpotential and ability to reverse the CO poisoning of the initial steps of electrolysis, has attracted attention largely on the basis of its ability to reduce CO2 to hydrocarbons. In order to overcome the problem of the poor solubility of CO2 in water, which represents a limiting factor for achieving acceptable current densities in preparative-scale electrolyses, Kaneko et al. [84] investigated the behavior of a system based on copper cathodes with methanol at 243 K as solvent, and with potassium or rubidium hydroxides as electrolytes. The reported yields on ethylene were higher than were obtained under similar conditions but in aqueous systems. In these studies, the authors also compared (in aqueous and methanol systems) the behavior of catholytes containing Rb or K to that of cations of a smaller radius (e.g., Li+), which were adsorbed more weakly to the cathode surface due to their strong hydration. The larger cations allowed much better yields in ethylene, conﬁrming the hypothesis that a low H+ coverage of the cathode, derived from the greater adsorption of large ions such as K+ or Rb+, favored the coupling of CH2: intermediates, whereas a high H+ coverage favored the reduction of a single intermediate to methane. The absence of any experimental data on the amount of charge passed makes it difﬁcult to evaluate any concrete possibilities of exploiting this approach in a scaled-up investigation. In fact, many attempts at using copper cathodes to reduce CO2 to hydrocarbons have focused on reducing the loss of selectivity that occurs due to the circulation of a limited amount of charge, as a result of poisoning species being formed on the cathode surface. To overcome this problem, Yano et al. [85] used a pulse mode of electroreduction, but extended the potential sweep of the pulse up to anodic values. In this way, the surface was periodically renewed, such that the accumulation of poisoning species on the surface was prevented. Positive results obtained with longer-range experiments (up to 30 min of pulse electrolysis) showing no change in selectivity towards hydrocarbon production suggested that this method could indeed be tested for scaled-up. How the experimental panorama is inﬂuenced by parameters still to be deﬁned was demonstrated by Shibata et al. [86]. Here, preliminary results obtained in aqueous media using a speciﬁc brand of high-purity commercial copper cathode were positive with regards to hydrocarbons C3+, provided that no electropolishing was performed before the electrochemical process. If electropolishing preceded the CO2 reduction, the cathodes behaved similarly to any other copper cathode, leading essentially (besides hydrogen) only to methane end ethylene. A tentative explanation of this behavior was proposed which referred to the polycrystalline matrix of this brand of copper, which made it particularly adaptable to be covered by oxide layers active in the formation of C3+. However, further experimental evidence on the surface structure, composition and modiﬁcations with electrolysis time will be required to substantiate this hypothesis. Other metals, the performances of which as cathodes in the electroreduction of CO2 have already been described in detail, were suggested for the development stage. Thus, Delacourt et al. [87] proposed an electrochemical cell, the design of which was taken from proton exchange membrane fuel cell (PEMFC) technology,

12.3 The Electroreduction of Carbon Dioxide in Protic Media (Water and Alcohols)

but suitably modiﬁed to attain better conductivities; silver cathodes, the catalytic activity of which in addressing the reduction towards the production of CO was well known, were also used. The aim of the study was to verify the suitability of such a cell design for the coproduction of CO and H2 to be used as syn gas in chemical plants. Long-term tests conﬁrmed the gradual onset, with duration of electrolysis, of a signiﬁcant worsening in selectivity that shifted towards the reduction of water to hydrogen. Tin has been recognized as an effective catalyst for the electroreduction of CO2. As with other metals with a high hydrogen overpotential (e.g., In, Pb, Hg, Cd), tin shows a very good selectivity and, under suitable conditions, produces formate in very high yields. Thus, Li and Oloman [88] have investigated the development of a continuous reactor at atmospheric pressure with a cathode constructed from a copper net onto which tin had been electrodeposited. Experiments with this threedimensional cathode on the inﬂuence of process variables on yields and selectivity towards formate provided promising results, such that the system was scaled-up and subsequently monitored initially in a small pilot plant [89]. The most signiﬁcant innovation – that of widening the range of products to longchain saturated hydrocarbons – was reported by Centi et al., who used carbon-based cathodes onto which Pt nanoparticles had been dispersed [90]. Remarkably, in the photoelectrocatalytic device the reduction side was in contact with a gaseous mixture of H2O and CO2. The product distribution, which ranged from methane to C8–C9 and possibly higher saturated hydrocarbons, appeared to be rather effectively inﬂuenced by the operating conditions and by the current density. In fact, the initial results did not allow the determination of current yields, nor of any precise product distribution, as a signiﬁcant part of the heavier products had remained adsorbed onto the porous carbon cathode. The authors noted that the ratios of their hydrocarbons yield did not ﬁt the expected (Schultz–Flory) product distribution of the Fisher–Tropsch synthesis. By comparison, Shibata et al. [86] claimed, over copper, a Schultz–Flory like product distribution, with yields at atmospheric pressure that were one order of magnitude higher than those reported by Centi et al. [90]. Interestingly, several of these groups suggested practical applications for their respective systems. For example, Centi et al. [90] proposed the use of their device in Mars missions, while Li and Oloman suggested that it be used in the industrial production of formic acid as an intermediate in industrial syntheses [88, 89]. Delacourt et al. considered the possibility of developing their production of syngas for methanol synthesis [87], while Kaneko proposed the selective reduction of CO2 to ethylene, as one of the most important building blocks of contemporary petrol chemistry [84]. An intriguing suggestion was made by Gattrel et al. [8], who compared the different pathways involving the reduction of CO2 through electrochemical processes fed by renewable energy; such a proposal might be considered more valid if it could be applied to the electrochemical enrichment of biogases. Unfortunately, as the current data were obtained from relatively short-term electrolyses, the scheme must ﬁrst be tested in the long term, so as to identify any practical problems that might arise due to the presence of impurities in the electrolytic system.

329

330

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

Acknowledgments

These studies were supported by the Ministero dell’Istruzione, dell’Universita e della Ricerca (MIUR) and the University of Palermo.

References 1 von Arnim, L.A. (1801) Annalen Der Physik, 8, 257–283. The Author, a known poet, writes a hymn to the unlimited power of the new energy source, illustrated with some experiments. Among them, the electrolysis of a French wine with the production of carbon dioxide was reported. Rather curiously, in the same journal, a few months in advance, another scientist, J.K.P. Grimm (7th band, 348–362) reported on the electrolysis of Hungarian wine, but no CO2 evolution was described. 2 Beketov, N.N. (1869) Russ. Phys. Chem. Soc., 1, 33–34. 3 Royer, M.E. and Hebd, C.R. (1870) Séances Acad. Sci. Fr. (Sect. Chimie), 70, 731. 4 Kolbe, H. and Schmitt, R. (1861) Ann. Chem., 119, 251–253. 5 Aresta, M. and Dibenedetto, A. (2007) Dalton Trans., 2975–2992. 6 Silvestri, G. (1987) Carbon Dioxide as a Source of Carbon: Chemical and Biochemical Uses (eds M. Aresta and G. Forti), Nato ASI Series, Ser. C, Riedel, Dordrecht, p. 339. 7 Silvestri, G., Gambino, S., and Filardo, G. (1990) Enzymatic and Model Carboxylation and Reduction Reactions for Carbon Dioxide Utilization (eds M. Aresta and J.V. Schlos), Kluwer, The Netherlands, pp. 101–127 and references cited therein. 8 Gattrel, M., Gupta, N., and Co, A. (2006) J. Electroanal. Chem., 594, 1–19. 9 Steckhan, E. (2002) Electrochemistry – Ullmann’s Encyclopedia of Industrial Chemistry, 7th edn, Wiley-VCH Verlag GmbH, Weinheim, Electronic Release. 10 (a) Silvestri, G., Gambino, S., and Filardo, G. (1991) Acta Chem. Scand., 45, 987–992.

11

12

13

14

15

16

17

18

19

20

(b) Sánchez-Sánchez, C.M., Montiel, V., Tryk, D.A., Aldaz, A., and Fujishima, A. (2001) Pure Appl. Chem., 73, 1917–1927. Silvestri, G., Gambino, S., Filardo, G., and Gullotta, A. (1984) Angew. Chem., 23, 979–980. Sock, O., Troupel, M., and Perichon, J. (1985) Tetrahedron Lett., 25, 1509–1512. Isse, A.A., Scialdone, O., Galia, A., and Gennaro, A. (2005) J. Electroanal. Chem., 585, 220–229. Pletcher, D., and Slevin, L. (1996) J. Chem. Soc., Perkin Trans. 2, 2, 217–220. Savéant, J.-M. (2006) Elements of Molecular and Biomolecular Electrochemistry: An Electrochemical Approach to Electron Transfer Chemistry, John Wiley & Sons, Inc. Scialdone, O., Filardo, G., Galia, A., Mantione, D., and Silvestri, G. (1999) Acta Chem. Scand., 53, 800–806. Isse, A.A., De Giusti, A., Gennaro, A., Falciola, L., and Mussini, P.R. (2006) Electrochim. Acta, 51, 4596–4964. Chaussard, J. (1990) Electrosyntheses from Laboratory, to Pilot, to Production (eds J.D. Genders and D. Pletcher), The Electrosynthesis Company Inc., New York, Ch 8. Scialdone, O., Belﬁore, C., Filardo, G., Galia, A., and Silvestri, G. (2004) Ind. Eng. Chem. Res., 43, 5006–5014. (a) Aishah, J., Hartini, M.A., Normala, S., Norhuda, A.M., Hanis, H.H.N., Razif, H.M., and Sugeng, T. (2007) J. Nat. Gas Chem., 16, 273–277. (b) Sigacheva, V.L., Frolovskii, V.A., and Petrosyan, V.A. (2005) Russ. J. Electrochem., 41, 1008–1014. (c) Chu, D.B., Li, X.H., Liu, X.Y., and Yao, W.L. (2004) Chin. J. Chem., 22, 1231–1234.

References

21

22

23

24

25

26

(d) Golinske, D., Voss, J., and Adiwidjaja, G. (2000) Coll. Czech. Chem. Comm., 65, 862–880. (e) Yamauchi, Y., Fukuhara, T., Hara, S., and Senboku, H. (2008) Synlett, 428–442. (f) Chiozza, E., Desigaud, M., Greiner, J., and Duñach, E. (1998) Stud. Surf. Sci. Catal., 114, 213–218. (g) Raju, R.R., Lateef, S.K., Mohan, S.R.K., and Reddy, S.R.J. (2006) Arkivoc, 76–80. (h) Ramesh, R.R., Mohan, K.S., and Reddy, J.S. (2003) Tetrahedron Lett., 44, 4133–4135. (i) Titov, V.E., Lopushanskaya, V.A., and Koshechko, V.G. (2002) Theor. Exp. Chem., 38, 289–294. (j) Koshechko, V.G., Titov, V.E., and Lopushanskaya, V.A. (2002) Electr. Commun., 4, 655–658. (k) Koshechko, V.G., Titov, V.E., and Lopushanskaya, V.A. (2000) Russ. J. Electr., 36, 149–154. Scialdone, O., Galia, A., Silvestri, G., Amatore, C., Thouin, L., and Verpeaux, J.N. (2006) Chem. Eur. J., 12, 7433–7447. (a) Fauvarque, J.F., Jutand, A., Francois, M., and Petit, M.A. (1988) J. Appl. Electrochem., 18, 109–115. (b) Favaurque, J.F., De Zelicourt, Y., Amatore, C., and Jutand, A. (1990) J. Appl. Electrochem., 20, 338. (a) Osakada, K., Sato, R., and Yamamoto, T. (1994) Organometallics, 13, 4645. (b) Olivero, S. and Dunach, E. (1999) Eur. J. Org. Chem., 8, 1885–1891. Gennaro, A., Isse, A.A., and Maran, F. (2001) J. Electroanal. Chem., 507, 124–134. (a) Amatore, C. and Jutand, A. (1991) J. Electroanal. Chem., 306, 141. (b) Amatore, C. and Jutand, A. (1991) J. Am. Chem. Soc., 113, 2819–2825. (a) Amatore, C., Jutand, A., Khalil, F., and Nielsen, M.F. (1992) J. Am. Chem. Soc., 114, 7086–7085. (b) Torii, S., Tanaka, H., Hamatani, T., Morisaki, K., Jutand, A., Pﬂuger, F., and Fauvarque, J.F. (1986) Chem. Lett., 169. (c) Damodar, J., Mohan, S.R.K., and Reddy, S.R.J. (2001) Electrochem. Commun., 3, 762–766.

27 (a) Isse, A.A., Gennaro, A., and Vianello, E. (1996) J. Chem. Soc., Dalton Trans., 1613–1618. (b) Damodar, J., Mohan, S.K., Lateef, S.K.K., and Reddy, S.J. (2005) Synth. Commun., 35, 1143–1150. 28 Zheng, G., Stradiotto, M., and Li, L. (1998) J. Electroanal. Chem., 453, 79–88. 29 (a) Simonet, J., Michel, M.A., and Lund, H. (1975) Acta Chem. Scand., B29, 489. (b) Lund, H. and Simonet, J. (1975) J. Electroanal. Chem., 65, 205–218. (c) Amatore, C., Pebay, C., Scialdone, O., Szunerits, S., and Thouin, L. (2001) Chem. Eur. J., 7, 2933–2939. 30 Isse, A.A., Ferlin, M.G., and Gennaro, A. (2003) J. Electroanal. Chem., 541, 93–101. 31 Isse, A.A. and Gennaro, A. (2002) Chem. Commun., 23, 2798–2799. 32 Isse, A.A., Ferlin, M.G., and Gennaro, A. (2005) J. Electroanal. Chem., 581, 38–45. 33 Scialdone, O., Galia, A., Errante, G., Isse, A.A., Gennaro, A., and Filardo, G. (2008) Electrochim. Acta, 53, 2514–2528. 34 Gennaro, A., Sánchez-Sánchez, C.M., Isse, A.A., and Montiel V. (2004) Electrochem. Commun., 6, 627–631. 35 Scialdone, O., Galia, A., Filardo, G., Isse, A.A., and Gennaro, A. (2008) Electrochim. Acta, 54, 634–642. 36 (a) Niu, D.F., Xiao, L.P., Zhang, A.J., Zhang, G.R., Tan, Q.Y., and Lu, J.X. (2008) Tetrahedron, 64, 10517–10520. (b) Olloqui-Sariego, J.L., Molina, V.M., González-Arjona, D., Roldán, E., and Domínguez, M. (2008) J. Electrochem. Soc., 155, E157–E161. 37 Scialdone, O., Galia, A., Filardo, G., Isse, A.A., Gennaro, A., Thouin, L., and Amatore, C. (2007) Giornate dell’Eletrochimica Italiana, Proceedings, Cagliari, Italy. 38 (a) Feroci, M., Orsini, M., Palombi, L., Sotgiu, G., Colapietro, M., and Inesi, A. (2004) J. Org. Chem., 69, 487–494. (b) Feroci, M., Inesi, A., Orsini, M., and Palombi, L. (2002) Org. Lett., 4, 2617–2620. 39 (a) Tokuda, M. (1999) Electrochemistry, 67, 993. (b) Senboku, H. and Tokuda, M. (2002) Fine Chem., 31, 50–60.

331

332

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation

40

41

42 43 44

45

46

47

48

49

(c) Chanfreau, S., Cognet, P., Camy, S., and Condoret, J.S. (2008) J. Supercrit. Fluids, 46, 156–162. (d) Mazin, V.M., Mysof, E.I., Sterlin, S.R., and Grinberg, V.A. (1998) J. Fuorine Chem., 88, 29–35. Anandhakumar, S., Sripriya, R., Chandrasekaran, M., Govindu, S., and Noel, M. (2009) J. Appl. Electrochem., 39, 463–465. (a) Stepanov, A.A., Volodin, Y.Y., Grachev, M.K., Kurochkina, G.I., Syrtsev, A.N., and Grinberg, V.A. (2002) Rus. J. Electrochem., 38, 1346–1352. (b) Stepanov, A.A., Grachev, M.K., Kurochkina, G.I., Sipin, S.V., and Grinberg, V.A. (2007) Russ. J. Electrochem., 43, 1211–1218. Isse, A.A. and Gennaro, A. (2003) Czech. Chem. Commun., 68, 1379–1394. Ikeda, Y. and Manda, E. (1985) Bull. Chem. Soc. Jpn, 58, 1723. Silvestri, G., Maspero, F., Gambino, S., and Filardo, G. (1992) Modern Methodologies in Organic Synthesis. Proceedings of International Symposium on Organic Reactions (ed. T. Shono), Kodansha, Tokyo, p. 44. (a) Silvestri, G., Gambino, S., and Filardo, G. (1987) US Patent 4 708 78. (b) Maspero, F., Piccolo, O., Romano, U., and Gambino, S. (1988) European Patent Application 26 944. (a) Chan, A.S.C., Huang, T.T., Wagenknecht, J.H., Miller, R.E. (1995) J. Org. Chem., 60, 742–744. (b) Datta, A.K., Marron, P.A., King, C.J.H., and Wagenknecht, J.H. (1998) J. Appl. Electrochem., 28, 569–577. Scialdone, O., Sabatino, M.A., Belﬁore, C., Galia, A., Paternostro, M.P., and Filardo, G. (2006) Electrochim. Acta, 51, 3500–3505. (a) Lateef, S.K., Raju, R.R., Mohan, M.S., Jayarama, R.S. (2006) Synth. Commun., 36, 31. (b) Damodar, J., Mohan, S.R.K., and Reddy, S.R.J. (2002) Synthesis, 3, 399–402. (a) Isse, A.A., Galia, A., Belﬁore, C., Silvestri, G., and Gennaro, A. (2002) J. Electroanal. Chem., 526, 41–52. (b) Scialdone, O., Galia, A., La Rocca, C., and Filardo, G. (2005) Electrochim. Acta, 50, 3231–3242.

50 (a) Jensen, H. and Daasbjerg, K. (1998) Acta Chem. Scand., 52, 1151– 1164. (b) Nadjo, L. and Saveant, J.M. (1971) J. Electroanal. Chem., 30, 41. (c) Lamy, E., Nadjo, L., Saveant, J.M.J. (1974) Electroanal. Chem., 50, 141–145. 51 Scialdone, O., Amatore, C., Galia, A., and Filardo, G. (2006) J. Electroanal. Chem., 592, 163–174. 52 Scialdone, O., Galia, A., Isse, A.A., Gennaro, A., Sabato, M.A., Leone, R., and Filardo, G. (2007) J. Electroanal. Chem., 609, 8–16. 53 Zhang, L., Xiao, L.P., Niu, D.F., Luo, Y.W., and Lu, J.X. (2008) Chin. J. Chem., 26, 35–38. 54 Wang, K.Z., Zhao, S.-F., Niu, D.-F., and Lu, J.-X. (2009) J. Electroanal. Chem., 630, 35–41. 55 Wang, H., Zhang, G.-R., Liu, Y.-Z., Luo, Y.-W., and Lu, J.-X. (2007) Electrochem. Commun., 9, 2235–2239. 56 Wang, H., Lin, M.-Y., Fang, F.-J., Chen, T.-T., and Lu, J.-X. (2007) Chin. J. Chem., 25, 913–916. 57 Grinberg, V.A., Koch, T.A., Mazin, V.M., Mysov, E.I., and Sterlin, S.R. (1999) Russ. Chem. Bull., 48, 295–300. 58 (a) Dinjus, E. (2001) Appl. Organomet. Chem., 15, 135–140. (b) Dénen, S., Duñach, E., and Périchon, J. (1991) J. Am. Chem. Soc., 113, 8447–8454. 59 Yuan, G.Q., Jiang, H.F., and Lin, C. (2008) Tetrahedron, 64, 5866–5872. 60 Koster, F., Dinjus, E., and Duñach, E. (2001) Eur. J. Org. Chem., 2507–2511. 61 Doherty, A.P. (2002) Electrochim. Acta, 47, 2963–2967. 62 Senboku, H., Yamauchi, Y., Fukuhara, T., and Hara, S. (2006) Electrochemistry, 74, 612–614. 63 Otero, M.D., Batanero, B., and Barba, F. (2006) Tetrahedron Lett., 47, 8215–8216. 64 (a) Wang, H., Zhang, K., Liu, Y.-Z., Lin, M.-Y., and Lu, J.-X. (2008) Tetrahedron, 64, 314–318. (b) Wang, H., Lin, M.Y., Zhang, K., Li, S.-J., and Lu, J.-X. (2008) Aust. J. Chem., 61, 526–530. 65 Orsini, M., Feroci, M., Sotgiu, G., and Inesi, A. (2005) Org. Biomol. Chem., 3, 1202–1208.

References 66 (a) Tascedda, P. and Dunach, E. (1995) J. Chem. Soc. Chem. Commun., 43–44. (b) Tascedda, P. and Duñach, E. (2000) Synlett, 245–247. (c) Tascedda, P., Weidmann, M., Dinjus, E., and Dunach, E. (2001) Appl. Organometal. Chem., 15, 113–120. (d) Casadei, M.A., Inesi, A., and Rossi, L. (1997) Tetrahedron Lett., 38, 3565–3568. (e) Casadei, M.A., Cesa, S., Feroci, M., and Inesi, A. (1999) New J. Chem., 23, 433–436. (f) Casadei, M.A., Cesa, S., and Rossi, L. (2000) Eur. J. Org. Chem., 10, 2445–2448. (g) Inui, T. (2001) Appl. Organom. Chem., 15, 141–144. 67 Yang, H.Z., Gu, Y.L., Deng, Y.Q., and Shi, F. (2002) Chem. Commun., 274–275. 68 Wang, H., Zhang, G.-R., Liu, Y.-Z., Luo, Y.-W., and Lu, J.-X. (2007) Electrochem. Commun., 9, 2235–2239. 69 Isse, A.A. and Gennaro, A. (2002) J. Electrochem. Soc., 149, D113–D117. 70 (a) Tyssee, D.A. (1976) US Patent 3.945.896. (b) Batanero, B., Barba, F., SánchezSánchez, C.M., and Aldaz, A. (2004) J. Org. Chem., 69, 2423–2426. (c) Scialdone, O., Sabatino, M.A., Galia, A., Filardo, G., and Silvestri, G. (2007) J. Electroanal. Chem., 614, 175–178. 71 Otero, M.D., Batanero, B., and Barba, F. (2006) Tetrahedron Lett., 47, 2171–2173. 72 (a) Feroci, M., Casadei, M.A., Orsini, M., Palombi, L., and Inesi, A. (2003) J. Org. Chem., 68, 1548–1554. (b) Rossi, L., Feroci, M., Verdecchia, M., and Inesi, A. (2005) Lett. Org. Chem., 2, 731–733. (c) Rossi, L., Feroci, M., and Inesi, A. (2005) Mini-Rev. Org. Chem., 2, 79–90. 73 (a) Tokuda, M., Kabuki, T., Katoh, Y., and Suginome, H. (1995) Tetrahedron Lett., 36, 3345–3348. (b) Tokuda, M., Yoshikawa, A., Suginome, H., and Senboku, H. (1997) Synthesis, 10, 1143–1145. 74 (a) Kamekawa, H., Senboku, H., and Tokuda, M. (1997) Electrochim. Acta, 42, 2117–2123. (b) Kamekawa, H., Kudoh, H., Senboku, H., and Tokuda, M. (1997) Chem. Lett., 917–918.

75

76

77 78

79 80

81

82

(c) Kuang, C., Yang, Q., Senboku, H., and Tokuda, M. (2005) Chem. Lett., 34, 528–529. Tokuda, M., Kabuki, T., and Suginome, H. (1994) Denkikagaku (presently Electrochem.), 62, 1144. (a) Kamekawa, H., Senboku, H., and Tokuda, M. (1998) Tetrahedron Lett., 39, 1591–1594. (b) Senboku, H., Fujimura, Y., Kamekawa, H., and Tokuda, M. (2000) Electrochim. Acta, 45, 2995–3003. (c) Senboku, H., Kanaya, H., Fujimura, Y., and Tokuda, M. (2001) J. Electroanal. Chem., 507, 82–88. (d) Senboku, H., Kanaya, H., and Tokuda, M. (2002) Synlett, 140–142. Tokuda, M. (2006) J. Nat. Gas Chem., 15, 275–281. Grinberg, V.A., Koch, T.A., Mazin, V.M., Mysov, E.I., and Sterlin, S.R. (1997) Russ. Chem. Bull., 46, 1560–1564. Bandeira, M.C.E. and Maia, G. (2008) Electrochim. Acta, 53, 4512–4519. Hori, Y. (2003) Handbook of Fuel Cells – Fundamentals, Technology and Applications, Volume 2: Electrocatalysis (eds W. Viestlich, H.A. Gasteger, and A. Lamm), John Wiley & Sons, Inc., pp. 720–733 (a) Reufer, C., Hateley, M., Lehmann, T., Weckbecker, C., Sanzenbacher, R., and Bilz, J. (2008) US Patent 7 332 067 Degussa. (b) Reufer, C., Hateley, M., Lehmann, T., Weckbecker, C., Sanzenbacher, R., and Bilz, J. (2008) US Patent EP16317027 Degussa. (c) Lehmann, T., Schneider, R., Weckbecker, C., Dunach, E., and Olivero, S. (2002) US Patent 2202/0053521 A1. (d) Rudisill, D.E., Panzone, G., Krauss, R.C., Strom, R.M., Scortichini, L.C., Kruper, W.J., Wolf, R.A., and Hay, D.A. (2001) US Patent 6 242 606, Merrell Pharmaceuticals Inc. (a) Rondinini, S., Mussini, P.R., Crippa, F., and Sello, G. (2000) Electrochem. Commun., 2, 491–496. (b) Kotsinaris, A., Kyriacou, G., and Lambrou, C. (1998) J. Appl. Electrochem., 28, 613–616.

333

334

12 Recent Scientiﬁc and Technological Developments in Electrochemical Carboxylation (c) Fedurco, M., Sartoretti, C.J., and Augustynnski, J. (2001) Langmuir, 17, 2380–2387. (d) Simonet, J. (2005) Electrochem. Commun., 7, 74–80. 83 Sànchez-Sànchez, G.M., Montiel, V., Tryk, D.A., Aldaz, A., and Fujishima, A. (2001) Pure Appl. Chem., 73, 1917–1927. 84 Kaneco, S., Katsumata, H., Suzuki, T., and Ohta, K. (2006) Electrochim. Acta, 51, 3316–3321. 85 Yano, J., Morita, T., Shimano, K., Nagami, Y., and Yamasaki, S. (2007) J. Solid State Electrochem., 11, 554–557.

86 Shibata, H., Moulijn, J.A., and Mul, G. (2008) Catal. Lett., 123, 186–192. 87 Delacourt, C., Ridgway, P.L., Kerr, J.B., and Newman, J. (2008) J. Electrochem. Soc., 155, B42–B49. 88 Li, H. and Oloman, C. (2006) J. Appl. Electrochem., 36, 1105–1115. 89 Li, H. and Oloman, C. (2007) J. Appl. Electrochem., 37, 1107–1117. 90 Centi, G., Perathoner, S., Wine, G., and Gangeri, M. (2007) Green Chem., 9, 671–678.

335

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass Michele Aresta and Angela Dibenedetto

13.1 Introduction

“Nature makes and chemists re-shape.” Among the industrial uses of carbon dioxide (CO2), the “enhanced biological ﬁxation,” which corresponds to the production of terrestrial or aquatic biomass under “non-natural photosynthetic conditions,” is increasingly attracting the attention of the scientiﬁc and technological world [1]. Such a process is responsible for life on Earth, as it concurrently converts inorganic carbon (Ci) – that is, CO2 or its hydrated form hydrogen carbonate, HCO3− – into organic compounds, and also generates dioxygen that is necessary for the life of living organisms. Such conversion of Ci into organics can occur either under natural conditions – that is, via the uptake of CO2 from the atmosphere, where it reaches a concentration equal to 0.038% (v/v) – or under “enhanced” or “industrial” conditions, that are much different from natural conditions. Typical examples of the “enhanced” biological ﬁxation are: (i) the cultivation of terrestrial biomass (ornamental plants, some vegetables) in greenhouses under a CO2 concentration in the gas phase of approximately 600 ppm; and (ii) the farming of aquatic biomass by dissolving CO2 in water or under a gas-phase concentration up to 5–10% – that is, 130- to 260-fold the natural concentration. Carbon dioxide ﬁxation is catalyzed by the enzyme ribulosebisphosphatecarboxylase oxygenase (RuBisCO), which is the most abundant natural enzyme worldwide, and the least selective. In fact, RuBisCO does not follow the usual selectivity of enzymes, and at the same time promotes the carboxylation of ribulose (a C-5 sugar) to afford a C-6 sugar (with carbon ﬁxation) and the oxidation of the same C-5 sugar, with a selectivity close to 50% [2]. Much effort has been spent on the genetic manipulation of RuBisCO, with the intention of increasing its selectivity towards the carboxylation reaction [3]. As a matter of fact, an improvement in carboxylation by 5–10% would solve all problems relevant to the accumulation of CO2 in the atmosphere, as all of the anthropogenic CO2 – which is causing the increases in atmospheric CO2 Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

336

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass

concentrations – would be ﬁxed into biomass. Yet, after more than two decades it is still not quite clear whether such biotechnology would have any possibility of implementation on a large scale. Nonetheless, ﬁxation under forced conditions remains the most usable “industrial” approach to biological CO2 ﬁxation. On the other hand, an intelligent exploitation of biomass could activate a “virtuous cycle” that would produce chemicals or energy compounds on a close-to zero-emission basis, with a positive effect on the reduction of CO2 emissions into the atmosphere. The potential of biomass utilization to control the accumulation of CO2 in the atmosphere forms the subject of this chapter.

13.2 The Natural Carbon Cycle

The natural carbon cycle is represented schematically in Figure 13.1. The total amount of carbon cycled per year ranges around 200 GtC, including ﬁxation into the terrestrial biomass (of any type), and in the watery environments. In aquatic media, CO2 is converted into biomass and inorganic carbonates upon reaction with water:

39 14 6

74

43

470 90 31

25

90

31 25 4200

Surface 900

150

FOSSIL FUEL

TERRESTRIAL BIOSPHERE

residential in GtC;

Intermediate 6600 Deep

Older soils 1500 inactive

580

31800

OCEAN

fluxes in GtC per year

Figure 13.1 The natural carbon cycle. The squares represent ﬁxed amounts of carbon in units of gigatons (Gt); the circles represent ﬂuxes in GtC per year. To convert to relevant amounts as CO2 ﬁgures, all values must be multiplied by 3.67.

13.3 The Utilization of Terrestrial Biomass

CO2 + 2 H2O = HCO3− + H3O+

(13.1)

2 HCO + Ca → CaCO3 + CO2 + H2O

(13.2)

− 3

2+

Noteworthy, the formation of hydrogencarbonate (Equation 13.1) is the process in which CO2 is ﬁxed, while the separation of calcium carbonate (Equation 13.2) causes the release of half of the ﬁxed CO2. The emission of CO2 from anthropogenic activities (the combustion of C-based fossil fuels, deforestation, combustion of woods) amounts to approximately 7.5 GtC per year, or about 3.5% of the total amount cycled in the natural cycle. However, as the natural systems are unable to use such CO2, this leads to its accumulation into the atmosphere. The assumption that an increase of the concentration of CO2 in the atmosphere would have boosted both the photosynthesis and the dissolution into the oceans has not been proven to be true. In fact, the solubility of CO2 is governed by complex equilibria, while photosynthetic ﬁxation is limited by several factors so that, under the increase of the atmospheric concentration from 280 ppm of the preindustrial era to the present-day 380 ppm, there has not been any sensible improvement of the uptake. Therefore, under natural conditions the uptake of CO2 has reached an equilibrium state, and the further increase in atmospheric concentrations may more likely cause climate changes through the “greenhouse effect” and destabilization of the thermal structure of the atmosphere, than improve the elimination of CO2 from the atmosphere. Consequently, there is a clear need to develop technologies for the efﬁcient sequestration of CO2. One such approach is the enhanced ﬁxation into biomass and the subsequent use of such biomass either for energy purposes or for making chemicals, thus avoiding the further extraction of fossil carbon. The production of energy and chemicals from biomass is expected to reduce CO2 emissions because the emitted CO2 will be again ﬁxed into the biomass: the effect will be that of expanding the natural carbon cycle. In order to win this game, it is necessary to produce and process biomass with a positive net total energy balance: namely, the energy accumulated into and extracted from a given amount of biomass must be higher than the energy input in the phases of production and processing of the same amount of biomass: Net Energy Production = Energy of the extracted fuel = Energy extracted from the biomass and ready for use − Energy consumed for the production and processing of that biomass The higher such difference, the more performing is the process. A process with a negative energy balance has no sense, as such a process is a net emitter of CO2.

13.3 The Utilization of Terrestrial Biomass

The terrestrial biomass has been used as source of energy since man lit the ﬁrst ﬁre on Earth. The direct combustion of any form of biomass is not the best process

337

338

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass

for energy production for several reasons, including the emission of toxic compounds derived from a noncomplete combustion of N-species, of NOx generated from the N-compounds present in the biomass, and other species that may have a large environmental impact. So far, energy has been derived from various types of biomass, and processes have been developed to keep the environmental impact as low as possible, and the energetic yield as high as possible. In general, the used biomass is either a residual from a variety of productive processes, or is cultivated for energy-creating purposes. 13.3.1 Residual Biomass

Several industrial processes produce residual biomass that may be used for energy purposes. One of the main sectors is the wood-industry, which produces considerable amounts of waste wood in the form of chips and pieces. Residual biomass is also produced in farming and agricultural activities, and/or in the agro-industry. All such types of biomass have a high content of cellulosic material and lignin; unfortunately, these polymeric materials are not easy to convert into other chemicals, and so in the past their most common use has been as co-fuels in thermal processes. However, this approach does not take advantage of the existence in such materials of complex chemical structures that often are synthesized industrially from fossil carbon, with large amounts of waste production and the use of environmentally nonbenign technologies. It would be much wiser to treat these materials in order ﬁrst to de-structure them, and then to recover the complex molecules that they contain, before adopting any technologies that lead to their destruction. (Table 13.1). Today, such approaches are under development, whereby technologies of different intensity can be used to make the best use of biomass as source of both chemicals and energy.

Table 13.1 The use of a “technologies cascade” for the full use of a biomass.

Very soft Nondestructive technologies

Soft Semi-destructive technologies

Hard Destructive technologies

Extraction of molecules with a complex molecular structure: molecular and polymeric compounds for special applications

Breaking of complex structures Production of energy products or simple chemicals: for example, lignin can be used for the synthesis of phenolic compounds upon hydrolytic treatment

Breaking of natural complex structures and production of very simple chemicals (CO–H2) that can be used for making again new complex molecular compounds (chemicals and fuels)

13.5 The New Generations of Biofuels

13.3.2 Cultivated Biomass

Cultivated biomass for energy purposes can be categorized as:

• • •

Cellulose-, hemicellulose- and lignin-rich biomass (e.g., forest trees, typically eucalyptus, poplar and similar fast-growing trees). Plants producing oil-rich seeds (e.g., sunﬂower, rapeseed, palm and many others). Plants producing starch-rich grains.

To date, these types of biomass have not been shown as the most proﬁtable sources of energy, although in some cases controversy has arisen that has resulted in negative attitudes towards the use of various types of biomass for energy production; these points will be discussed in the following sections.

13.4 The First-Generation Biofuels

The ﬁrst-generation biofuels can be identiﬁed as ethanol, which was produced via the alcoholic fermentation of cereals, and bio-oil or biodiesel, which was extracted from seeds such as sunﬂower, rapeseed, or palm. The use of cereals and sunﬂowers was rejected by public opinion and some scientiﬁc environments, because their use for energy production conﬂicted with their use as foodstuffs. In fact, the diversion of cereals to the production of ethanol for transport has led to a rise in the price of ﬂour and derived goods, especially in Mexico. The same situation has arisen for some bio-oils, such that the source was shifted to palm-oil which, essentially, is produced in Asian countries such as Malaysia. Consequently, the ﬁrst-generation ethanol, oils and bio-oils are no longer at the center of attention, and are not expected to contribute towards solving the problem of CO2 emissions reduction in the transport sector. It is more likely that there will be a major surge of interest to identify substitutes for these materials. Some possible solutions to these problems are discussed later in the chapter.

13.5 The New Generations of Biofuels

The new generations of biofuels must respond to the issues of reducing CO2 accumulation in the atmosphere through a quasi-zero-emission energy production, but without competing with the food market. Hence, biomass that has no nutritional value for either humans or animals must be taken into consideration.

339

340

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass

13.5.1 Second-Generation Biofuels

Typically, forest trees and nonedible seeds must be used for the production of fuels, although these require a new approach when compared to materials used to date with regards to their handling and energy recovery. Biomasses which are rich in cellulose, hemicellulose and lignin are currently under investigation, and new technologies are today being developed for separation (cellulose, lignin), depolymerization (cellulose, lignin), and fermentation (cellulose) methodologies that may produce fuels for the transportation sector, notably ethanol or biodiesel. The production of bio-oil for thermal uses is today also receiving much attention.

•

Cellulose: Currently, cellulose is replacing cereals for the production of ethanol. The depolymerization of cellulose represents one of the key issues for obtaining monomeric or low-oligomeric species that can easily be fermented so as to afford ethanol. The dissolution of cellulose in ionic liquids (ILs) [4] has provided great expectations, as the dissolved cellulose may easily be digested by bacteria to afford ethanol. The use of ILs is not perceived as a completely friendly technology, however, and as a consequence biodriven depolymerizations (using fungi or bacteria), combined with chemical technologies, is used more extensively. The process of cellulose degradation is shown schematically in Figure 13.2. The residual biomass and organic compounds derived from the process can be further processed to afford either biogas by anaerobic fermentation (CH4) or Syngas (CO + H2) by using a more destructive treatment such as a reaction in supercritical water or harsh thermal processes. Cellulose is also directly converted into syngas, which can be used for the synthesis of

Storage tank Storage tank, (cellulose) ethanol Blow tank Washer Distillation Enzymatic Digester Ethanol Screw press hydrolysis fermentor Screw press; hydr/fer Solid fuel m Storage tank Evaporator I (lignin) Product tank, hemicellulose

Evaporator IILignin processing Product tank, lignin Figure 13.2 Schematic representation of the treatment of polymeric materials as cellulose and conversion into energy products such as ethanol.

13.5 The New Generations of Biofuels

either methanol or gasoline and diesel. Syngas can also be further treated using the water gas shift reaction and converted into CO2 and H2 that can be easily separated [5] to afford hydrogen. Although cellulose can be processed using different technologies to afford a variety of fuels such as ethanol, gasoline, diesel, hydrogen, and methanol, this approach requires the development of new catalysts for biomass conversion or else syngas conversion. In fact, syngas generated from biomass has different properties from that produced from fossil carbon or hydrocarbons.

•

Lignin: The fragmentation of lignin through the cleavage of ethereal bonds may produce a wide variety of chemicals that have a complex structure (Figure 13.3).

The use of cellulose, hemicellulose and lignin may result in the production of large amounts of interesting chemicals with a large market (ranging from kilotons to megatons per year), but which would normally be generated from fossil carbon. This process would require major energy consumption and lead to high CO2 emissions, with a low selectivity, and likewise produce large amounts of waste that would eventually result in the generation of CO2. 13.5.2 Third-Generation Biofuels

Currently, aquatic biomass is recognized as a third-generation source of biofuels. However, the need to decouple energy issues from both land use and food production is resulting in the application of pressure towards exploiting aquatic biomasses such as microalgae, macroalgae, and plants. In fact, it is difﬁcult to envisage that the terrestrial biomass alone will be capable of producing the large volumes of biofuels required to meet a target of a 20% substitution of transport fossil fuels with biofuels by 2020, as has been proposed by several industrialized countries [6]. One reason why interest in aquatic biomass has developed so rapidly, and with such driving force, is that algae are better converters of solar energy (η = 6–8% under natural conditions, up to 9–10% in bioreactors) than are terrestrial plants (η = 1.5–2.2%), and also have a better potential for fuel production diversiﬁcation In addition, the composition of microalgae and macroalgae may be more easily addressed than that of superior plants, mainly because algae are much simpler organisms (unicellular). Compared to terrestrial annual seed plants that produce oil (e.g., sunﬂower, rapeseed, palm), algae are much more efﬁcient; as an example, the best case of oil yield for terrestrial plants will not exceed 10% of the whole plant, whereas algae can reach 75% (e.g., Botryococcus braunii, Schizodetrium sp.) with a good average of about 40–50% (Nannoclhoropsis sp., Nitzschia sp.), while macroalgae have an average productivity of 20% (Codium fragile, Cladophora). One consequence of this superior solar energy conversion is that aquatic cultures will have a bio-oil productivity of between 50 and 130 m3 ha−1 [7], compared

341

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass

342

H2COH

OH

H2COH

CH2

CH2

CH2 H2COH

CH2

OCH3

CH3

CH

CH2OH

HC

OH

O

HC

O H2COH

OCH3

O O

HC

O

HC OCH3

HC

HC O

OH H2C

HOCH2

HC O O CH2O

CH

CH

O

CH

CH

HC

H2COH

CH HCOH

CH3O

H2COH

CH

HC

HOCH2

HC

OCH3 O

CO

HC

CH

HC

CH2

CH3O

HC

HOCH2 O

CHO CH

O

OH

CH

OCH3 O

O

CO

O

CH3O

OCH3

CH3O

H2COH

CHO

CH2OH HC

O

HCOH

O

HOCH2 CH3O

O

HC

H2COH

CH3O

H2COH O

HOC

CH

CH

HOCH

CH

O

OCH3 O

H2COH

CH

CH3O

H 2C

HC

OH

OCH3

HOCH

H2COH

OCH3 CH3O

C2H CH

HCO

CH3O

(Carbohydrate)

HC

CH3O

CH

HO

CH

O

H2COH

CH2O

CH

CH3O

CH

CH3O

HCOH HOC

H2COH

HO

CH H2COH

O

CH2 HC

OCH3 O

HCOH

HCOH HCOH

CH3O O

H2CO

OCH3 OH

Figure 13.3

The structure of lignin, and the phenolic moieties that may be derived from it.

to a terrestrial biomass such as palm oil, which has a best production value of 6 m3 ha−1. As aquatic biomass can be grown on marginal coastal areas, on desert lands close to salty water, and/or off-shore, this greatly increases it potential as a source of fuel in comparison with their terrestrial counterparts. In the case of aquatic biomass, the main issues that require attention are the cultivation techniques, the harvesting technologies, and the ﬁnal processing of the algal product. Each of these procedures will require energy that would reduce to varying extents, depending on the type of biomass under consideration, and the net energy produced by the algae.

13.5 The New Generations of Biofuels Inﬂuence of the CO2 concentration on the distribution of fatty acids in Chaetomorpha L. cultured at ambient conditions and under a high concentration (10% in the gas-phase) of CO2.

Table 13.2

% CO2 in gas-phase

Fatty acid

Total FAMEs

14 : 0

16 : 0

16 : 1

18 : 0

18 : 1

18 : 2

20 : 0

20 : 4

20 : 5

Control 0.038

5.2 ± 1.7

9.4 ± 1.6

0.9 ± 0.2

0.5 ± 0.2

5.9 ± 1.8

5.9 ± 1.2

0.2 ± 0.1

0.5 ± 0.2

0.6 ± 0.2

29.1 ± 4.3

Enhanced 10.0

5.0 ± 1.1

16.2 ± 2.0

0.8 ± 0.1

0.5 ± 0.3

11.0 ± 1.6

19.2 ± 2.5

0.3 ± 0.1

1.1 ± 0.2

1.6 ± 0.4

55.5 ± 3.7

Microalgae can also be grown in bioreactors installed in many different formats, including ﬂat, vertical, slanting, or coiled. Typically, solar light or artiﬁcial white light would be employed for irradiation [8]. Algae have the ability to adapt quickly to speciﬁc growing conditions, and will also react strongly to any changes in these parameters. For example, the application of physical stress during the growth process can be used in a positive manner to increase the cellular content of lipids, proteins and other components; however, a negative factor here is that the culture management must be strictly controlled, requiring great care and expertise, if the quality of the biomass is to be maintained. For example, the responses of a macroalga to the gas-phase concentration of CO2 is shown in Table 13.2. Based on these data, it is evident that if an alga is grown under a gas-phase concentration of CO2 that is 260-fold that found naturally, then the cellular production of lipids and the fatty acid distribution will be greatly affected, with a clear shift towards a larger degree of unsaturation of the fatty acids [9]. Based on this ability to manipulate the algal composition, these organisms can be used for the production of different types of biofuel. For example, those algae which are rich in lipids are better suited for the production of bio-oil or biodiesel; those rich in starch can be used for alcoholic fermentations to afford ethanol; and those rich in proteins and starch can be used for the production of biogas. Unlike terrestrial plants, which may produce bio-oil with a high concentration of a single fatty acid (>90%), algal bio-oils may contain a wide variety of fatty acids, and these may in some way be addressed or controlled (Table 13.3). For example, olive oil contains triglycerides in which the glycerol is esteriﬁed mostly with oleic acid (56–85%), with palmitic acid (3–20%) and linoleic acid (7–20%) as the two other major components, whilst palm oil contains triglycerides of palmitic acid (46%) and oleic acid (38%), and soybean oil contains 50% linoleic acid (17–30% oleic acid, 9–13% palmitic acid). Table 13.4 shows that algae, unlike terrestrial plants, may produce lipids derived from a wide variety of fatty acids that may not be a positive fact, especially if several poly-unsaturated fatty acids are present. The quality of bio-oil or biodiesel extracted from algae closely resembles that of fuels

343

344

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass Table 13.3 Distribution of fatty acids in lipids present in some macroalgae. Values are relative

percentage of organic compounds. Fatty acid (no. of carbon atoms/no. of unsaturated bonds)

Macroalgal species Ulva lactuca

Enteromorpha compressa

Padiva pavonica

Laurencia obtuse

Saturated C12→C20

15.0

19.6

23.4

30.2

Mono-unsaturated C14→C20

18.7

12.3

25.8

9.0

Poly-unsaturated C16/2→C16/4 C18/2→C18/4, C20/2

66.3

68.1

50.8

60.9

Table 13.4 Low heating value (LHV) of biodiesel and bio-oil produced from biomass.

Origin

Biodiesel, LHV(MJ kg−1)

Bio-oil, LHV (MJ kg−1)

Sunﬂower seeds Jatropha Microalgae

36 34 36

12 13 12

extracted from terrestrial plants, based on the comparative low heating values(LHVs) of oils extracted from the two sources (Table 13.4). Both, microalgae and macroalgae are rich in chemical compounds that can be extracted by using a variety of different methods. Thus, by adopting a strategy involving a “cascade of technologies” it should be feasible to use aquatic biomass for the production of high-value chemicals and fuels. Such an approach would be particularly important as the lower costs of producing algae would in turn reduce the cost of algal biomass-derived fuel, making it more competitive with fuels obtained from fossil sources (see below). Compounds that can be extracted from microalgae and macroalgae include:

• • • • • •

coloring substances and antioxidants; enzymes (superoxide dismutase, restriction enzymes, phosphoglycerate kinase, luciferase, luciferin); polymers (polysaccharides, starch, poly-β-hydroxybutyric acid); peptides, toxins, amino acids, steroids, essential oils such as geraniol-geranyl formate or acetate-cytronellol-nonanol-eucalyptol. pigments, such as chlorophylls, carotenoids, xantophylls; and amines and inorganic compounds.

13.5 The New Generations of Biofuels

Such a wide product-entropy does not necessarily make the extraction of a given product economically convenient, however. Notably, the ability of algal organisms to concentrate certain types of material under conditions of stress may help to reduce the entropy and to increase the concentration of a product in the biomass. In fact, it is possible to direct cultures of algae towards the production of either chemicals (e.g., astaxanthine or carotenoids) or high-energy products (e.g., bio-oil or starch). The implementation of a technology cascade would allow the production not only of materials with a complex structures, but also high-energy products or more simple molecules such as H2 and CO (syngas), using the most appropriate technology. One issue that recently has attracted much attention is the transesteriﬁcation of bio-oil to produce glycerol. The molecular structures of lipid materials extracted from algae, notably glycerol and triglycerides, are shown in Scheme 13.1. The production of fatty acid methyl esters (FAME)s from lipids (Scheme 13.2) is based on the use of an aqueous solution of a base (NaOH) and methanol, to produce aqueous glycerol (bioglycerol) and the FAME. The glycerol may then be separated by distillation from the process solution, albeit with a large consumption of energy, and used in a variety of applications [10] (Figure 13.4). Whereas, in the past, the production of glycerol from fatty acids virtually satisﬁed global demand, such production has steadily increased during the past decade (Figure 13.5) to a point where it will shortly exceed market demand, although this is due also in part to the rapid expansion in biodiesel production (Figure 13.6). Hence, the identiﬁcation of a new market for glycerol [11] is of paramount importance, or this coproduct of biodiesel will go to waste. It is also becoming increasingly urgent to identify new transesteriﬁcation technologies for the production of nonaqueous glycerol, as well as new technologies for converting glycerol into products that may ﬁnd large applications, perhaps as

CH2OH

CH2OOC-R

I

I

CHOH

CHOOC-R

I

I

CH2OH

CH2OOC-R

Glycerol

Triglyceride

R: C8H17CH=CHC7H14 in oleic acid CH3(CH2)14 in palm acid CH3(CH2)7CH=CHCH2CH=CH(CH2)4 in linoleic acid

Scheme 13.1 Structures of glycerol, triglycerides (lipid or bio-oil) and fatty acids (FA).

CH2OOC-R CHOOC-R

CH2OH +

NaOH + 3CH3OH

CH2OOC-R Lipid

CHOH

+ 3 R-COOCH3

CH2OH Methanol

Scheme 13.2 Transesteriﬁcation of lipids.

Glycerol

FAME

345

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass

3%

17%

10%

Others Alkydic resins Cosmetics, soaps, medicine Paper Esterss Tobacco Polyglycerol Food and drinks Resale Cellulose

6%

9% 28% 12%

3%

Figure 13.4

11%

1%

Industrial applications of glycerol.

3000 2500

1000 tons

346

2000 1500 1000 500 0 1990

1995

2000

2005

2010

2015

year

Figure 13.5

World production of glycerol.

3000 2500 2000

others

1000 MT 1500

soaps fat alcohols

1000

fat acids biodiesel

500 0 1992

1995

1998

2001

2004

2007

2010

year

Figure 13.6

Sources of glycerol and the increasing production derived from biodiesel.

reaction intermediates [12], as additives for fuels [13], or in the production of H2 [14]. Today, bioglycerol chemistry represents a truly “hot topic.” In addition to creating bio-oil from algae, it is also possible to produce ethanol. Whilst the aerobic fermentation of sugars to produce alcohol has been recognized

13.6 Implementation of the Bioreﬁnery Concept

for millennia, aquatic biomasses also contain variable quantities of simple sugars, starch and cellulosic materials that are suited to the production of ethanol. Moreover, other alcohols (e.g., butanol) can be coproduced in smaller amounts. Today, the production of bioethanol from biomass, as an alternative to corn, constitutes a major area in the research of alternative fuels. Today, the anaerobic fermentation of fresh organic material is widely used in the production of biogas, which is a mixture of methane and CO2. This practice is implemented on a large scale, with fresh fruit vegetable garden (FGV) residues, such as household waste, being converted into biogas and a slurry. Whilst the efﬁciency of the process depends on several parameters, the quality of the biogas product depends on the feed and the operating conditions [15], with reactors of different types being used, together with mesophilic microorganisms and thermophilic bacteria [16]. Recently, one major application of this technology has been in the production of biogas by fermenting fresh municipal waste, which is converted into energy (methane) for use by the community. Based on their compositions, both algae and aquatic biomasses are well suited for conversion into biogas although, unfortunately, not all aquatic biomasses provide the best energy yield and some may prove to be inadequate for such use. Some species of macroalgae, however, have been shown to serve as excellent sources of biogas [17]. The production of hydrogen from an aquatic biomass can be achieved in several ways, the preferred sources for dihydrogen being sugars, acids, and other molecular compounds; any polymeric materials must be depolymerized before use. The hydrogen-producing bacteria used in fermentation broths for biogas production are highly active during the ﬁrst 8 h, which is termed the “acidiﬁcation phase” (see Refs [14, 18] and references therein). These bacteria are easily isolated and can be used to produce hydrogen from polyols and acids. Short-chain fatty acids such as acetic acid and lactic acid can be further converted into dihydrogen and CO2 by using photosynthetic bacteria such as Rhodospirillum rubrum [19]. Bioglycerol is particularly suited to the production of dihydrogen; in this case, bioglycerol produced during the transesteriﬁcation of lipids in aqueous solution can be used, without any need for puriﬁcation [14]. An alternative approach is to treat the biomass in supercritical water so to convert the biomass into syngas, which is a mixture of CO and H2 [20] that can be further converted into H2 and CO2 by using the water gas shift reaction (WGSR): CO + H2O → CO2 + H2

(13.3)

On completion of the reaction, the H2 and CO2 can be easily separated using a mature membrane.

13.6 Implementation of the Bioreﬁnery Concept

As noted above, an aquatic biomass contains a much wider variety of molecules when compared to fossil carbon or oil. Hence, the exploitation of their full

347

348

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass

potential must be carefully planned by deﬁning the most appropriate process technologies. The “bioreﬁnery” approach is the most sound in terms of truly exploiting the potential of an aquatic biomass, and this concept is now becoming accepted on a worldwide basis. In the bioreﬁnery approach, the economic and energetic value of the biomass is maximized, although it must be emphasized that ﬂuctuations in the prices of fossil carbon (coal, oil, gas) raises uncertainty regarding the opportunity to produce biodiesel from aquatic biomass. For example, when the oil price is below US$ 120 per barrel it is uneconomic to produce biodiesel in this way. On the other hand, an aquatic biomass demonstrates an excellent potential for use as a source of specialty chemicals, with some components also having added value as animal feeds or fertilizers. Integrating the production of chemicals and fuels may eventually lead to the production of biofuels as an economic process. Indeed, may recent reports have described aquatic biomass as having good potential for applying the concept of the bioreﬁnery, although examples which certify the beneﬁts of a value chain exploitation have not (yet) been provided. One key point in the development of algal cultures on a large scale is the decision to grow them in ponds or in photobioreactors (PBRs). Whilst the latter approach may decouple algal cultivation from climatic conditions and offer the advantage of controlled culture conditions, the technology employed is generally more energy- and materials-intensive, such that any beneﬁts are reduced. The location of the cultures is also a key issue; the use of arable land would be excluded, as it would involve the same concerns as when growing terrestrial biomass. Off-shore cultivation may provide an interesting solution, but has certain drawbacks due to the costs of management. The use of internal desert environments or marginal coastal areas may provide major economic beneﬁts for poorer regions and communities, while natural or artiﬁcial basins could be considered for the exploitation of waste water and resource recovery. As noted above, municipal and process waters can be used, with the additional beneﬁt of water cleaning and better resource utilization. Fisheries may represent interesting candidates for coupling water treatment and the growing of aquatic biomass for energy or chemicals production. In this case, the water recycling would reduce the ﬁsh production costs, while the further use of an algal biomass would reduce the energy costs of the aquafarm such that, eventually, it might become energy self-sufﬁcient. In order to take the most advantage of aquatic biomass, it is necessary to incorporate an integrated approach to its exploitation, by combining expertise in areas of algae cultivation, nanotechnology applications for processing the biomass with process intensiﬁcation, and the production of new materials from the aquatic biomass. The implementation of the bioreﬁnery concept is perhaps the most scientiﬁcally and technologically sound approach for the real exploitation of the aquatic biomass, and indeed such an attitude is now beginning to be accepted on a worldwide basis. Ultimately, it may contribute to providing a portfolio of products which may produce a positive global economic and energetic balance for growing aquatic biomass.

References

Taken together, whilst aquatic biomass represents an interesting source of chemicals and energy, additional accurate investigations must be conducted before its true potential is revealed.

13.7 Concluding Remarks

The bioreﬁnery concept brings to deﬁnition a production plan that allows all useful fractions of a biomass to be used. However, aquatic biomass represents a much wider variety of raw materials compared to fossil fuels, and their full potential must be exploited by deﬁning appropriate transformation routes and the most adequate technologies. The direct combustion of aquatic biomass is, indeed, the most primitive technology for the use of algae, and the less remunerating from an energetics point of view. It would also be the most polluting in terms of products emitted into the atmosphere. In the case of algae, the ratio (R): R = energy spent for producing biodiesel energy given back by the extracted biodiesel = 1.2 − 0.3 will depend on the species and strain of the alga under consideration. This means that a number of parameters (algal strains, growing and harvesting technologies, algae separation and drying technologies, treatment processes) must be very carefully selected and assessed; otherwise, the risk is that more energy will be spent than is produced. The efﬁciency of biofuel production is very important: a value of R close to 1 or higher than 1 will exclude the biomass as a candidate for energy production. The interesting aspect here is that an aquatic biomass can furnish not only chemicals but also energy products. These chemicals may have a high added value (as in the case of astaxanthine or pigments), such that the costs of algal production are covered; consequently, the process would be not only economically feasible but also energetically acceptable. The production of aquatic biomass focusing initially only on energy production may represent a risky operation, taking into consideration today’s large ﬂuctuations in the price of fossil-based oil. As noted above, with fossil-oil prices currently below US$ 120 per barrel, algal biodiesel is barely competitive with diesel from fossil fuels. However, if the oil price were to exceed US$ 120 per barrel, then biodiesel from aquatic biomass may become economically viable [21, 22].

References 1 Aresta, M. and Dibenedetto, A. (2009) in Handbook of Combustion (eds M. Lackner, F. Winter, and A. Agarwal), Wiley-VCH, Weinheim.

2 Aresta, M. and Forti, G. (eds) (1987) Carbon Dioxide as a Source of Carbon: Chemical and Biochemical Uses, Reidel Publ., NATO-ASI C-206.

349

350

13 Indirect Utilization of Carbon Dioxide: Utilization of Terrestrial and Aquatic Biomass 3 (a) Akazawa, T. (1987) Carbon Dioxide as a Source of Carbon: Chemical and Biochemical Uses (eds M. Aresta and G. Forti) Reidel Publ., NATO-ASI, pp. 83–91. (b) Parry, M.A.J., Andralojc, P.J., Mitchell, R.A.C., Madgwick, P.J., and Keys, A.J. (2003) J. Exp. Botany, 54 (386), 1321–1333. (c) Andersson, I. and Backlund, A. (2008) Plant Physiol. Biochem., 46 (3), 275–291. 4 (a) Kosan, B., Michels, C., and Meister, F. (2008) Cellulose, 15, 59–66. (b) Fukaya, Y., Hayashi, K., Wada, M., and Ohno, H. (2008) Green Chem., 10, 44–46. 5 (a) Feng, X. and Huang, R.Y.M. (1996) J. Membr. Sci., 116 (1), 67–76. (b) Allam, R.J., Bredesen, R., and Drioli, E. (2003) Carbon Dioxide Recovery and Utilisation (ed. M. Aresta), Kluwer Publishers. (c) Aresta, M., Dibenedetto, A., Pastore, C., and Fragale, C. (2008) ChemSusChem, 1, 742–745. 6 Target 20 20 20. http://ec.europa.eu/ environment/climat/climate_action.htm. 7 Chisti, Y. (2007) Biotechnol. Adv., 25 (3), 294–306. 8 (a) Ono, E. and Coello, J.L. (2003) Greenhouse Gas Control Technologies, Special Issue Dedicated to GHGT, Pergamon, pp. 1503–1508. (b) Ono, E. and Coello, J.L. (2007) Biosyst. Eng., 96, 129–134. 9 (a) Aresta, M., Dibenedetto, A., Carone, M., Colonna, T., and Fragale, C. (2005) Environ. Chem. Lett., 3, 136–139. (b) Aresta, M., Dibenedetto, A., and Barberio, G. (2004) Preprints -Am. Chem. Soc. Div. Fuel Chem., 49 (1), 348–350. (c) Aresta, M., Ribenedetto, A., and Barberio, G. (2005) Fuel Process. Technol., 86, 1679–1693. (d) Aresta, M., Ribenedetto, A., and Barberio, G. (2008) Algae World 08, 17–18 November, Swissotel Merchant Court Singapore. (e) Aresta, M. and Ribenedetto, A. (2008) DGMK Tagungsbericht, vol. 161, DGMK, ISBN: 978-3-936418-81-1. 10 (a) Medina A.R., (1999) J. Bioechnol., 70, 379–391.

11

12

13

14

15 16

17

18

(b) Roice, M., Subhashchandran, K.P., Gean, A.V., Franklin, J., and Rajasekharan Pillai, V.N. (2003) Polymer, 44, 911. (c) Wilson, R., van Schie, B.J., and Howes, D. (1998) Food Chem. Toxicol., 36, 711–718. (d) Garcia, R., Besson, M., and Gallezot, P. (1995) Appl. Catal. A: General, 127, 165–176. (e) Wittlich, P., Themann, A., and Vorlop, K.D. (2001) Biotechnol. Lett., 23, 463–466. (f) Clacens, J.M., Pouilloux, Y., and Barrault, J. (2002) Appl. Catal. A: General, 227, 181–190. Pagliaro, M., Ciriminna, R., Kimura, H., Rossi, M., and Della Pina, C. (2007) Angew. Chem., Int. Ed., 6 (24), pp. 4434–4440. Aresta, M., Dibenedetto, A., Ferragina, C., and Nocito, F. (2008) EU Patent 08305653.1-2117 to ARKEMA – FR. Kesling, J. Jr, Haven, S., Karas, J.L., Liotta, J.L. Jr, and Frank, J. (1994) US Patent 530836 to ARCO Chemical Technology, L.P. (Wilmington, DE). Aresta, M. and Dibenedetto, A. (2009) Catalysis for Sustainable Energy Production (eds P. Barbaro and C. Bianchini), Wiley-VCH, Weinheim, p. 444. Aresta, M., Narracci, M., and Tommasi, I. (2003) Chem. Ecol., 19, 451–459. (a) Aresta, M. and Galatola, M. (1997) Energy from Wastes (ed. Nowatech), Laterza, Bari, Italy. (b) Valorga (1985) The Valorga Process, Second Annual International Symposium on Industrial Resource Management, Philadelphia, 17–20 February. (c) De Baere, L. and Boelens, J. (1999) Organic Waste Systems NV, Dok Noord 4, 9000 GENT, ows.be (accessed 1998). (a) Samson, R. and Leduy, A. (1982) Biotechnol. Bioeng., 24 (8), 1919–1924. (b) Vergara-Fernández, A., Vargas, G., Alarcón, N., and Velasco, A. (2008) Biotechnol. Bioeng., 32 (4), 338–344. (a) Nath, K. and Das, D. (2003) Curr. Sci., 85 (3), 265–271. (b) Milne, T.A., Elam, C.C., and Evans, R.J. (2001) A Report for the International Energy Agency, IEA/H2/TR-02/001.

References 19 Melnicki, M.R., Bianchi, L., De Philippis, R., and Melis, A. (2008) Int. J. Hydrogen Energy, 33 (22), 6525–6534. 20 (a) Lv, P., Yuan, Z., Wu, C., Ma, L., Chen, Y., and Tsubaki, N. (2007) Energy Convers. Manag., 48 (4), 1132–1139. (b) Lightner, G.E. (2000) US Patent 6133328.

21 Benemann, J.R. (1997) Energy Convers. Manag., 38, S475–S479. 22 Sheehan, J., Dunahay, T., Benemann, J., and Roessler, P. (1998) A Look Back at the U.S. Department of Energy’s Aquatic Species Program – Biodiesel from Algae. National Renewable Energy Laboratory, Golden.

351

353

14 Fixation of Carbon Dioxide into Inorganic Carbonates: The Natural and Artiﬁcial “Weathering of Silicates” Ron Zevenhoven and Johan Fagerlund

In this chapter the mineralization of carbon dioxide (CO2) is addressed, from the viewpoints of: (i) valuable inorganic carbonate material production; and (ii) large-scale carbon capture and storage (CCS). Analogies with the natural weathering of rock and the material resources are discussed, followed by a summary of the state-of-the-art of carbonate production for the two different viewpoints mentioned.

14.1 Introduction: Inorganic Carbonate Uses and Natural Resources

Compared to organic carbonates that are produced by oxidative carboxylation of hydrocarbons using various homogeneously or heterogeneously catalyzed routes developed during the twentieth century, the use and production of inorganic carbonates has a very long history. Most important are the (earth) alkaline carbonates with limestone and soda as the best known examples. Construction works have, over several millennia, made use of cement and concrete, with the Egyptian pyramids and the Great Wall of China still standing. Typically, the use and production methods involve large volumes of low-value materials and cheap processing. For example, the hardening of cement is largely a matter of time, the presence of water, and the temperature and humidity of the local surroundings. Worldwide, cement production is of the order of 2–2.5 Gt (gigatons) per year, while the use of so-called “crushed rock”, mainly composed of limestone (calcium carbonate) is an order of magnitude larger still. Soda, as sodium carbonate (soda ash or washing soda), has been used in glass making for centuries and, similarly to sodium hydroxide (caustic soda), in processes where alkaline water solutions are used, such as papermaking or washing-type activities. Sodium bicarbonate on the other hand is applied as so-called baking soda or baking powder. Besides the (bi)carbonates of calcium and sodium, also those of potassium known as potash, or carbonate of potash, have been used since antiquity (when it was produced from wood ashes) as fertilizer, and in glass and soap production. The bicarbonate ﬁnds use as baking powder and several other applications, similar to sodium Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

354

14 Fixation of Carbon Dioxide into Inorganic Carbonates

bicarbonate. Much less extensive is the use of (bi)carbonates of magnesium, barium, lithium, and strontium (Mg, Ba, Li, Sr), that ﬁnd applications in rubber processing, and the production of glass, ceramics, photochemicals, cosmetics and medicine, catalysts, and batteries. Still, other carbonates that have been used in applications like pigments are for example lead carbonate (white lead) [1, 2]. Calcium carbonate can be mined as limestone, and sodium carbonate and sodium bicarbonate, be it at much smaller amounts, as soda, nahcolite, trona or carbonate hydrates. Calcium bicarbonate is water-soluble and found almost exclusively in aqueous solutions. While the limestone form of calcium carbonate can usually be found and mined not far from where needed, an extensive market exists for synthetic or precipitated calcium carbonate (PCC) (∼ 13 Mt per year worldwide in 2007). PCC is (mainly) used in the paper industry, plastics and rubber, and paint production, while a roughly ﬁvefold larger market exists for ground calcium carbonate (GCC) in these sectors [3]. The worldwide production of soda (ash), produced from trona or nahcolite mineral or sodium carbonate-containing brines was ∼45 Mt in 2007 [3]. Since the 1880s the industrial-scale production and upgrading of sodium (bi)carbonate has proceeded via the Solvay process, where basically carbon dioxide (CO2) is combined with sodium from seawater. The process uses ammonia (NH3) that is regenerated using lime (calcium oxide) that is produced from limestone, ﬁnally giving calcium chloride as byproduct [4]. Most of the worldwide soda ash production proceeds through Solvay-based plants. “Potash” is used as a collective name for mined or produced salts that contain water-soluble potassium (for a large part as sylvite, potassium chloride mineral), with a total worldwide production of ∼35 Mt in 2007 [1]. Potassium (bi)carbonate production methods involve the carbonation of potassium hydroxide (KOH), which in turn is produced by electrolysis of aqueous potassium chloride solutions. The carbonate of magnesium (magnesite) can be mined as such (∼15 Mt worldwide in 2007, some 45% of which is mined in China) or as dolomite (magnesium calcium carbonate), or it can be produced by carbonating a magnesium chloride brine. The production of carbonates of barium, strontium and lithium is typically achieved by carbonating barium sulfate (barite), strontium sulfate (celestine) and lithium brines (containing chloride, sulfate and (bi)carbonate) at typical amounts of 10–100 kt per year. The use of lead carbonate (produced from lead acetate and CO2) as pigment may be decreasing as for all nonbattery lead products; the processing of scrap lead-acid batteries may involve the carbonation (using sodium carbonate) of lead sulfate (see Ref. [5]). To conclude, carbonate materials are either mined from rock material or produced from natural brines. Often, the processing or upgrading involves a carbonation step, which then means that some CO2 is taken out of a gas stream that otherwise would be emitted. At the same time, carbonate material may be produced by carbonating an industrial waste or byproduct from a process where, presumably, CO2 from the same process can be used. These two options of CO2 ﬁxation, aiming either at CO2 emissions reduction or carbonate production, will be further addressed below.

14.2 Natural Fixation of CO2 in Carbonates

14.2 Natural Fixation of CO2 in Carbonates

Many technical–chemical processes take maximum beneﬁt of similarities with ongoing processes in Nature, with increased purity or reaction speed as the most important differences. The production of carbonates is a typical example of this, and the process of CO2 mineralization for carbon capture and storage (CCS) (see Section 14.4) is in fact the accelerated version of what is known as the “natural weathering” of minerals. This is a combination of the interacting processes of mechanical and chemical weathering, and relevant to the current discussions are the chemical weathering processes of dissolution and hydrolysis that involve CO2 [6, 7]. A dissolution equilibrium reaction that proceeds in Nature with dissolved CO2 in water and calcite gives a bicarbonate solution: CO2(g ) + H2O ( l ) + CaCO3(s) ⇔ Ca2+ (aq ) + 2 HCO3− (aq )

(14.1)

which is in fact the reaction responsible for the dissolution of coral structures. Similarly, also other carbonates will to some extent dissolve in water. Rau and coworkers suggested an application of this for large-scale CCS in an accelerated weathering of limestone (AWL) plant, using waste limestone ﬁnes from limestone processing. The product would be a calcium bicarbonate solution that could be readily released and diluted into the ocean with a minimal or even a positive environmental impact [8, 9]. However, further research is needed before this alternative can be applied at any larger scale, as there are still many issues to deal with, such as the energy demand of transporting large amounts of calcium containing (waste or mineral) material to the AWL plant that preferably should be located near a CO2 point source, as well as a possible disposal site (e.g., the ocean). In an ideal case (with access to free limestone, such as waste ﬁnes, and a “free” water source, such as power plant cooling water) the CO2 mitigation cost by means of AWL could be as low as 3–4 $ ton−1 CO2 (2.3–3.1 € ton−1 CO2). It was suggested that some 10–20% of the United States point-source CO2 emissions could be mitigated in this way. The environmental effects of bicarbonate solution disposal into the ocean were also discussed: while direct CO2 injection into the ocean lowers the pH, the release of a bicarbonate calcium ion-containing solution could actually counteract the ongoing ocean pH reduction. In order to avoid negative impacts to the ocean, the CO2-containing ﬂue gas should be free of impurities such as heavy metals. Despite the potential positive effect of bicarbonate disposal, it is concluded that further research is required to fully understand the impacts of AWL efﬂuent disposal in the ocean [9]. An approach similar to that discussed above is the carbonation of Ca-containing waste materials by creating a reaction surface and simply spraying water onto it; this simple, low-cost approach might even be applied to larger-scale systems for capturing CO2 from ambient air in the future [10]. The amounts of CO2 that could be sequestered using Ca-based materials are, however, not large enough for a signiﬁcant CCS impact. However, when using naturally occurring magnesium

355

356

14 Fixation of Carbon Dioxide into Inorganic Carbonates

silicates this could be an option [11], although this would require a more aggressive weathering, namely hydrolysis. The hydrolysis of silicate minerals involves chemical reactions with weakly acidic or alkaline aqueous solutions. With natural waters that contain dissolved CO2 – that is, CO2(aq) which can be written as carbonic acid (H2CO3) – the hydrolysis of magnesium silicate (olivine) proceeds as 4 CO2( g ) + 4 H2O ( l ) + Mg2SiO4 (s) ⇔ 2 Mg2+ (aq ) + 4 HCO3− (aq ) + 2 H4SiO4 (aq ) (14.2) giving silicic acid (which may form a silica gel) and dissolved magnesium bicarbonate. A similar process can also be applied to the dissolution of wollastonite (CaSiO3) and larnite (Ca2SiO4). Another important example is potassium feldspar weathering that produces kaolinite: 2 CO2(g ) + 11 H2O ( l ) + 2 KAlSi3O8 (s) ⇔ Al 2Si 2O5(OH)4 (s) + 2 K + (aq ) + 2 HCO3− (aq ) + 4 H4SiO4 (aq )

(14.3)

A third chemical weathering mechanism that is of importance is oxidation/ reduction that involves mainly the elements carbon, iron, manganese and, of course, oxygen. An equilibrium reaction between dissolved CO2 and bicarbonate ions can lead to the precipitation of ferrous iron, giving a hematite (ferric oxide) precipitate: 1 2 Fe2+ (aq ) + 4 HCO3− ( aq ) + O2( g ) ⇔ 4 CO2( g ) + 2 H2O ( l ) + Fe2O3(s) (14.4) 2 This oxidation of iron is of importance for carbonation of magnesium oxide-based silicates for CCS, using for example olivine which a mixture of Mg2SiO4 (forsterite) and Fe2SiO4 (fayalite). In other similar situations, the oxidation of Fe or Mn ions can limit the reaction with (dissolved) CO2. This short analysis shows that dissolved CO2 can “attack” a wide range of minerals and rocks, resulting in the release of all sorts of cations and binding of CO2 as dissolved bicarbonate or carbonate ions, effectively preventing it from being released to the atmosphere. Furthermore, depending on the pH and the saturation of the solution, solid carbonates (and in some cases also bicarbonates) can precipitate, removing it from the hydrosphere as well. With enormous amounts of water, CO2 and rock material available on the Earth, the scale on which the above occurs in Nature is enormous. Together with the ﬁxation of CO2 that produces biomass and the natural weathering of rocks, the CO2–bicarbonate–carbonate cycle forms an important part of the Earth’s entire carbon cycle. From a CCS point of view, it also means that CO2 mineralization technology taken into use on a large scale has it parallels in Nature. Along the same lines, it was reported very recently from China [12] and India [13] that silicate and carbonate rock weathering in rivers gives rise to increased concentrations of dissolved ions such a Mg2+, Ca2+, Na2+, HCO3− , and Cl−, and that the CO2 consumption of this should be taken into consideration as a “sink” for “greenhouse gases.” Further details on CO2 reactions in aqueous solutions can be found elsewhere [6, 14].

14.3 Process Routes to Valuable Carbonate Products

14.3 Process Routes to Valuable Carbonate Products

The most important inorganic carbonate materials, their natural resources, and the conventional process routes were listed in Section 14.1. When the goal shifts from carbonate material production to a process that reduces CO2 emissions, or ﬁxes signiﬁcant amounts of CO2, then different process routes are followed, mainly because the raw materials are different. It is important for the discussion below to distinguish between direct and indirect process routes. Direct carbonation is the simplest approach to carbonate production (or mineral carbonation; see Section 14.4) and the principal approach is that a suitable feedstock – for example, serpentine or a Ca/Mg-rich solid residue – is carbonated in a single process step. For an aqueous process this means that both the extraction of metals from the feedstock and the subsequent reaction with the dissolved CO2 to form carbonates takes place in the same reactor. If, on the other hand, the process of mineral carbonation is divided into several steps, it is classiﬁed as indirect carbonation. In other words, indirect carbonation means that the reactive component (usually Mg or Ca) is extracted from the feedstock (typically as oxide or hydroxide) in one step and then, in another step, it is reacted with CO2 to form the desired carbonates. 14.3.1 Material Resources

Industrial solid residues that contain large amounts of Mg and Ca (and even Fe) are of great interest for carbonation, as they can combine in an economically feasible manner beneﬁts of: (i) reducing the overall CO2 emissions; (ii) producing a valuable carbonate material; and/or (iii) improving the quality of the industrial residue for disposal. Since around the year 2000, the carbonation option has been of research interest, offering better economic prospects than large-scale CO2 mineralization for CCS. The most important materials are listed in Table 14.1 (note also the four technology reviews that have been produced so far [15–18]). The carbonation of fractions of industrial byproducts and residues or consumer wastes is largely limited to calcium-containing waste streams, where often the production of high-value calcium carbonates such as PCC is aimed at. Many calcium-containing industrial residues have an unstable nature and potentially high reactivity, especially when compared to virgin rock material. Treating waste products with CO2 has the possibility of rendering for example heavy metals immobile [19, 20]. Industrial residues or byproducts that have been studied for the purpose of carbonation include: asbestos-mining tailings, electric arc furnace (EAF) dust, steel-making slag [10, 21–27], waste concrete [10, 28], cement-kiln dust [29, 30], coal ﬂy ash [31], air pollution control (APC) residues [32, 33], municipal waste incinerator (MSWI) ash [34, 35], pulverized fuel (PF) ﬁring and circulating ﬂuidized-bed combustion (CFBC) ashes [36] and oil shale ashes [37], as well as ash transportation waters [38].

357

358

14 Fixation of Carbon Dioxide into Inorganic Carbonates Table 14.1 Industrial residue material resources for carbonation.

Industry sector

Material

Iron and steelmaking

Steel slag (general); blast furnace slag; steel converter slag; AOD process slag; electric arc furnace slag

Cement industry

Waste cement; cement kiln dust

Waste incineration

APC slag; MSWI bottom ash

Power production

APC slag; pulverized coal ﬂy ash; lignite ﬂy ash; CFBC ash; oil shale ash

Paper industry

Paper bottom ash

Mining industry

Mining waste/tailings

Other

Contaminated soil; brines

The most important among these appear to be steelmaking slags and cement industry wastes, for which the costs of CO2 ﬁxation where estimated at US$ 8 ton−1 CO2 [10]. For year 2003, it was estimated that the worldwide productions of iron slags and steel slags were 160–200 Mt and 96–145 Mt, respectively, with 1.4 Mt produced in Finland annually [21]. With a more recent estimate of 315–420 Mt per year for iron- and steelmaking slags produced worldwide, this gives a total CO2 ﬁxation potential of 60–170 Mt per year [39]. For cement waste, the number given for Japan was 123 Mt cement waste produced in 2001, which could be used to ﬁx 30 Mt CO2, corresponding to ∼10% of the country’s CO2 emissions for that year (312 Mt) [40]. Cement waste carbonation may be of considerable importance for major cement-producing countries such as China and India, both having rapidly expanding economies with CO2 emissions that are increasing accordingly. For MSWI bottom ashes and APC residues, Costa et al. [41] reported the state-of-the-art three years ago, pointing out that conclusions as to speciﬁc metal species (Pb, Zn, Cu) stabilization cannot be produced without great caution. More recently, the team of Baciocchi compared wet (100% CO2, 303–323 K, 0.3 MPa) and dry routes (10% CO2, 673 K, 0.1 MPa) for MSWI APC residue carbonation, and found a maximum conversion of ∼65% in a matter of minutes for both routes, ﬁxing ∼250 g CO2 kg−1 APC residue [42]. For Europe, the CO2 storage potential of this material was <0.05% of the CO2 produced in the region. 14.3.2 Direct (Single-Step) Process Routes

The gas–solid carbonation of calcium-based materials is being investigated for the separation and concentration of CO2 using calcium oxide/calcium carbonate carbonation/calcination cycles (e.g., Ref. [43]), but not for the production of valuable carbonate materials.

14.3 Process Routes to Valuable Carbonate Products

Similarly, the wet, aqueous processes can also be used to improve the stability of ashes from waste or solid fuel combustion. In some cases, such as Estonian oil shale the ashes bind signiﬁcant amounts of CO2, often allowing for simple and cheap processing. On the other hand, the amounts of solid material will not be such that an effect noticeable from a CCS point of view is achieved, while at the same time the produced carbonate cannot be qualiﬁed as a valuable product. 14.3.3 Indirect (Single-Step) Process Routes 14.3.3.1 General Aspects of Calcium Carbonate Production The production of valuable carbonate products implies in practice the production of a valuable (precipitated) calcium carbonate. When producing these from calcium-containing waste materials, several process parameters have a direct inﬂuence on product quantity and quality. Many test results and product analyses have been reported, mainly for low-value feedstock materials such a cement waste, ashes and slags from iron- and steel production, and the carbonation processes are practically all based on aqueous systems operating at conditions up to 473 K, 20 MPa. Based on the required product purity and quality, processes can be divided into:

•

An apparently single step process: here, calcium (Ca) is extracted immediately before the carbonation; that is, CO2 gas is present during this step and the carbonation can take place right-away.

•

A two step process: here, the extraction of calcium is followed, after separation, by carbonation of this, presumably under different process conditions.

While the apparently single-step process partly involves gas–solid carbonation chemistry, the two-step process implies carbonation of dissolved calcium only, and is primarily a crystallization process. In order to achieve a valuable carbonate product, the single-step conversion is generally followed by a product improvement step, for example a recrystallization. Therefore, this is referred to here as an apparent single step. Based on reports in the open literature (e.g., Refs [19, 26, 44–49]), the following can be concluded with regards to product amounts and properties:

•

The CO2 pressure increases the carbonation rate, especially the initial carbonation rate, while the overall carbonation result (% conversion of Ca to CaCO3) stays the same.

•

The temperature primarily affects how much product is obtained. For a given CO2 pressure, a certain optimum temperature exists for a certain feedstock, its particle size (distribution), and the liquid/solid (L/S) mass ratio (kg kg−1) used. However, the L/S ratio is not a very sensitive parameter, with typical values of the order of 10.

•

The temperature optimum for a given set of other parameters is due to low leaching rates of calcium at too-low temperatures, and decreasing activities of (bi)carbonate ions at too-high temperatures.

359

360

14 Fixation of Carbon Dioxide into Inorganic Carbonates

•

The pressure of CO2 will affect the pH of the solutions, with one outcome that higher pH levels will result in smaller particles. Likewise, also mixing and stirring will affect particle size (distribution), to a large extent following general experiences with crystallization. Strong relationships between the product particle size (distribution) and the process conditions do not exist, however.

•

When considering particle size and its distribution, typically the particle size D3,2 (the average diameter if all particles were to have the same outer surface) is more important than D4,3 (the average diameter if all particles were to have the same volume), which in turn is more important than a D50 average diameter, which then again is more important than the speciﬁc surface area (SA) (m2 g−1). For a certain threshold particle size, the results obtained can be much better than for larger particles.

•

For a given CO2 pressure and particle size (distribution), the effect of temperature (which determines yield) can be small; in that case, the use of an additive is an option. When considering costs it is essential that such additives can be recovered and recycled.

•

Considering pressure, a threshold value is found above which the amount of dissolved CO2 and (bi)carbonate, together with the dissolved calcium (determined also by L/S ratio) exceeds the calcium carbonate solubility product, such that precipitation can occur.

•

Widely used additives as chlorides (NaCl, KCl) and/or bicarbonates (NaHCO3, KHCO3), which affect the solubility of Ca and other species, increase ionic strengths and increase the activity of the (bi-)carbonate ions. Sugar-type additives have also been used. Phosphor in the feedstock ends up less in the product when bicarbonate additives are used.

•

The preheating of feedstock material increases the Ca-release rates and amounts.

•

Other species such as Mn and Cr, have a strong (usually negative) impact on product whiteness.

•

Increasing the L/S ratio results in more Ca being extracted, while trace elements and heavy metals leaching is not affected to any great extent.

•

For the separation of Fe and species such as Al, Cr, Mn, and Ti, a magnetic separation may be beneﬁcial; however, signiﬁcant amounts (perhaps one-third) of the calcium present may then also be removed (as 2 CaO·Fe2O3).

•

For the leaching and carbonation of calcium, the enzyme carbonate anhydrase may be used to catalyze the chemistry. A careful control of pH and dissolved metallic species is required, however.

•

Preferable feedstocks have a low Fe content, a high Ca/Si ratio, and a particle size <0.5 mm.

14.3 Process Routes to Valuable Carbonate Products

•

361

The chlorine content need not be a problem, as this may give an increased ionic strength.

14.3.3.2 Acetic Acid Route In order to speed up the aqueous carbonation process, the use of acetic acid for the extraction of calcium from a calcium-rich feedstock has been suggested by Kakizawa et al. [50]. In principle, this process comprises two steps, as given in Equations 14.5 and 14.6 [50]:

CaSiO3 + 2 CH3COOH → Ca2+ + 2 CH3COO− + H2O + SiO2

(14.5)

Ca + 2 CH3COO + CO2 + H2O → CaCO3 + 2 CH3COOH

(14.6)

2+

−

In principle, the acetic acid used in the extraction step could be recovered in the following precipitation step. The process is illustrated in Figure 14.1. Teir et al. [51, 52] expanded the idea presented by Kakizawa et al. [50] to PCC, and showed that its production via the conventional route will give CO2 emissions of the order of 0.21 kg kg−1 PCC (assuming oil combustion for lime calcination), whereas PCC production via the acetic acid route using wollastonite implies a net ﬁxation of 0.34 kg CO2 kg−1 PCC. [2] (see also Ref. [53]). In practice, complete recycling of the acid/extraction agent is necessary for a process to become feasible on a large scale, but this has not yet been demonstrated in a cost and energetically effective manner. Inspired by the concept of binding CO2 in calcium extracted from a calcium silicate such as wollastonite using acetic acid [50], Teir et al. [52, 54] investigated the possibility of producing a high-value PCC material from calcium silicates. Later, the concept was developed further in order to identify other calcium-containing materials that would replace the relatively expensive calcium silicate source wollastonite [21, 55]. Steelmaking slags are currently the center of attention, as they can contain signiﬁcant amounts of both CaO and MgO, where ammonium salts are used to select calcium selectively from the slag, while being fully recoverable if ammonia vapor losses can be prevented (see also Ref. [56]).

CH3COOH

Extraction CH3COOH

Separation

Carbonation

Separation

Ca2+ + 2CH3COO + CO2 + H2O =>

CaSiO3 + 2CH3COOH => Ca2+ + 2CH3COO + H2O + SiO2

CaCO3 ↓+ 2CH3COOH

CaSiO3

SiO2

CO2

Figure 14.1 The precipitated calcium carbonate (PCC) production process: A schematic representation, according to the acetic acid route [21, 51].

CaCO2

362

14 Fixation of Carbon Dioxide into Inorganic Carbonates

Although, globally, the CO2 sequestration potential for this option is small, for individual steel plants he method could provide signiﬁcant economical beneﬁts and noticeable reductions in emissions. 14.3.3.3 Two-Step Aqueous Carbonation of Solid Residues By upgrading a waste product into a product of high commercial value, expensive CO2 ﬁxation processes could become economically feasible. One such approach has been investigated by Katsuyama et al. [57], who studied the use of waste cement for the development of high-purity CaCO3 by CO2 carbonization in accordance with the process scheme in Figure 14.2. Katsuyama et al. [57] studied the feasibility of producing CaCO3 from waste cement by ﬁrst extracting calcium from pulverized waste cement in a water slurry at high CO2 pressures (several MPa), followed by the precipitation of CaCO3 from the extracted solution at lower CO2 pressures, producing high-purity CaCO3 (up to 98%) from waste cement at relatively high reaction rates. It was estimated that the cost of producing high-purity CaCO3 could be as low as US$ 136 m−3 (€105 m−3), which compares to the commercial price of US$ 200–350 m−3 (€154–269 m−3). In addition, if the produced CaCO3 could be puriﬁed to meet the requirements of ultrahigh purity CaCO3 (>99% CaCO3) the potential proﬁts could be increased substantially. The current cost of ultrahigh purity CaCO3 is approximately US$ 10 000 m−3 (€7700 m−3), while Katsuyama et al. [57] estimated a production cost of only US$ 323 m−3 (€250 m−3). Much along the same lines as the process described above, as proposed by Katsuyama et al. [57], Geerlings et al. recently described a process for producing CaCO3 from various solid residues [58]. In this patent, two examples of the described process were described, one utilizing paper bottom ash and one steel slag, although neither example provided a reaction rate for the precipitation step.

Waste concrete

Pulverizer Pulverizer

Recycle aggregate

Classifier Classifier

Compressor

Waste cement particle

Building demolition

CO2 separation

CO2 Water Residue

Extraction Extraction reactor reactor Desulfurization Desulfurization Ca-rich water Precipitation Precipitation reactor reactor

Figure 14.2

Flue gas CaCO3

Schematic illustration of a CaCO3 production process from waste cement [57].

14.3 Process Routes to Valuable Carbonate Products

The extraction of calcium took place inside a water-ﬁlled, stirred reactor for 15 min, which resulted in a calcium hydroxide concentration of 1.1 g l−1 for paper bottom ash and 0.46 g l−1 for steel slag. The formed hydroxide slurry was separated from the solids and carbonated by injecting it with CO2 at a rate of 25 ml min−1. However, the feasibility of the process should be investigated with a cost and environmental assessment before any ﬁrm conclusions can be drawn, as was also the case for a process described by Gorset et al. [59]. Gorset et al. [59], when describing a method of producing pure MgCO3 from olivine, claimed that the process, which consisted of one dissolution step and two precipitation steps (both MgCO3 and amorphous silica), was fast enough for largescale implementation (dissolution rates ca. 1.5 × 10−12 mol cm−2·s). The process did not require the use of strong mineral or organic acids, even though the dissolution step required an acidic environment. The required acidity (pH 3–5) was achieved using pressurized CO2 (5–15 MPa) and a temperature of about 373–433 K, while the following step of MgCO3 precipitation took place in another reactor, preferably with a lower CO2 pressure (5–8 MPa) and a higher temperature (413–523 K) favoring the precipitation of carbonates. The experimental results showed a high degree of purity, between 99.28 and 99.44% MgCO3, of the precipitated carbonate. 14.3.3.4 The pH-Swing Process The pH-swing process, which was developed in Japan (and later also presented in a patent by Yogo et al. [60]) is another two-step aqueous carbonation process where at ﬁrst the pH of the solution is lowered so as to enhance the extraction of divalent metal ions. In the second step, the pH is raised to enhance the precipitation of carbonates. A schematic representation of a process utilizing the pHswing is shown in Figure 14.3 (taken from Ref. [61]), where the principal reactions

Boiler exhaust (CO2: 13 %)

NH3 + CaCl2 (basic solution)

CO CO22 absorber absorber CO2 absorbed gas

(1) Ca extraction (2) CO2 absorption / CaCO3 precipitation Figure 14.3

NH4Cl (acidic solution)

(NH4)2CO3 + CaCl2

Alkaline-earth metal source (2CaO·SiO2)

(2) (2) Precipitator Precipitator

(1) (1) Extractor Extractor

CaCO3

Extraction residue

Flow diagram of the pH-swing process [61].

363

364

14 Fixation of Carbon Dioxide into Inorganic Carbonates

taking place inside the extractor (Equation 14.7) and the precipitator (Equation 14.8) are: 4 NH4 Cl + 2 CaO⋅SiO2 → 2 CaCl 2 + 4 NH3 + 2 H2O

(14.7)

4 NH3 + 2 CO2 + 2 H2O + 2 CaCl 2 → 2 CaCO3 + 4 NH4 Cl

(14.8)

Equation 14.8, which takes place inside the precipitator, consists of both CO2 absorption and CaCO3 precipitation. In their study, Kodama et al. [61] investigated a CO2 sequestration process that utilizes pH-swing using NH4Cl. The energy input requirement for the investigated process using steel-making slag as a mineral source was estimated at approximately 300 kW·h t−1 CO2, but the loss of a chemical additive (NH3) was considerable. Investigations on the NH3 loss problem are ongoing [61].

14.4 Mineral Carbonation for Carbon Capture and Storage (CCS)

In the ﬁeld of the long-term storage of CO2, mineral sequestration is a CCS option that provides an alternative for the more widely advocated method of geological storage in underground cavities, especially at locations where such underground cavities are not available, where the risk of leakage of the CO2 stored underground is considered unacceptable, or where large resources of material suitable for carbonation are present. Although the technical state-of-the-art of mineral carbonation processing suffers from too-slow chemical kinetics and poor energy economy, the driving forces for continued attention for this CCS route are its sheer capacity (dwarﬁng other CCS methods), the fact that it gives compact and leakage-free CO2 ﬁxation that needs no post-storage monitoring, and ﬁnally the potential of operating at a zero (or negative!) net energy input, provided that the process is properly optimized, taking beneﬁt of favorable thermodynamics. It is important to distinguish between in situ and ex situ approaches. In situ mineral carbonation is closely connected to the underground storage option, as it involves the injection of CO2 into underground reservoirs. The difference is that, in situ mineral carbonation explicitly aims at reacting the CO2 to form carbonates with alkaline-minerals present in the geological formation. In a recent report [62], the advantages of in situ mineral carbonation compared to ex situ processes were highlighted for basaltic rock carbonation as applied to at Hellisheidi, Iceland within the CarbFix project, where the mass of rock to be moved in ex situ processes is considered to make it “impractical”. Similar to the natural weathering processes, the overall reaction between a metal oxide-bearing material and CO2 can be expressed as: MO + CO2 ↔ MCO3 + heat

(14.9)

where in practice M describes a (metallic) element such as calcium, magnesium, or iron. The reaction in Equation 14.9 is exothermic, and the heat released is

14.4 Mineral Carbonation for Carbon Capture and Storage (CCS)

dependent on the metallic element-bearing mineral at hand (for the magnesiumor calcium-based silicate minerals olivine: 89 kJ mol−1 CO2, serpentine: 64 kJ mol−1 CO2 and wollastonite: 90 kJ mol−1 CO2 at 298 K) [63] One major beneﬁt of CO2 sequestration by mineral carbonation consists of the environmentally benign and virtually permanent trapping of CO2 in the form of carbonated minerals, by using abundant mineral resources such as Mg-silicates [63]. Moreover, unlike other CO2 sequestration routes, it provides a leakage-free long-term sequestration option, without a need for post-storage surveillance and monitoring once the CO2 has been ﬁxed. When Teir et al. [64] investigated the stability of calcium and magnesium carbonate when subjected to an acidic aqueous environment (such as acidic rain water), they concluded that Ca/Mg carbonates should be resistant enough to prevent local environmental effects at a mineral carbonate storage site, although a slight dissolution in water could be beneﬁcial as dissolved HCO3− would give an alkalinity buffer against acidiﬁcation, while the Mg2+ ions would support biomass growth. In addition to these beneﬁts of mineral carbonation, this option is the only CO2 sequestration option available where large underground reservoirs do not exist, and/or ocean storage of CO2 is out of the question, such examples being Finland [65] and Korea [66]. In Lithuania, and in the Baltic region in general [67], alternatives to in situ CO2 trapping are being explored, as the saline aquifers in Lithuania have been found unsuitable for CO2 storage [68]. In Norway, an important olivine (and other minerals) producer and a leader in CCS technology demonstration, the mineralization of CO2 was also initiated [69, 70]. Yet another very important beneﬁt of mineral carbonation is that, at least in theory, the carbonation process could proceed without energy input, although this has not yet been accomplished in practice. In fact, many of the carbonation processes presented to date have suffered from being too energy demanding and expensive. The biggest challenge has perhaps been – and still is – to enhance the otherwise extremely slow (hundred thousands of years in Nature) carbonation reaction, without incurring excessive overall process costs. Attempts to speed up the carbonation reaction have included the use of both dry and wet methods, additives, heating and pressurizing the carbonation reactor, dividing the process into multiple steps, pretreatment of the mineral source, and more. Nonetheless, although a steadily increasing number of research teams continue to investigate these methods, none of them has proven to be both economically and environmentally viable, due to various difﬁculties related to mineral carbonation. As a result, according to IEA (2008), “It is unlikely that mineralization will offer an opportunity for sequestering large volumes of CO2” [71]. When considering the reaction temperature, an increase in the temperature of a process is known to enhance the reaction rate. However, thermodynamics places restraints on the stability of carbonates, such that the temperature can be increased only to a certain (pressure-dependent) level before the formation of CO2 is favored over carbonates. For example, at CO2 pressures of 0.1 MPa, MgCO3 is stable up to temperatures of about 673 K, whereas at 3.5 MPa CO2 it is stable up to about 823 K [2]. Therefore, a simple solution of increasing the temperature (and making use

365

366

14 Fixation of Carbon Dioxide into Inorganic Carbonates

of the exothermic high-temperature carbonation reaction) until the reaction rates were sufﬁciently fast would not work without pressurization; hence, other (indirect) routes must be investigated. Another factor to be considered is that the large-scale sequestration of CO2 from ﬂue gases as mineral carbonates will require vast amounts of mineral: typically, 1 kg of CO2 may require 2 kg (or more) of serpentine for disposal, and this would certainly result in a signiﬁcant environmental impact at the disposal site [2]. On the other hand, a process for storing several Mt of CO2 per annum would involve solids handling of a scale similar to a typical metal ore or mineral mining and processing activity; some examples are given in Ref. [18]. An example of a very large-scale processing of solids is the mining and processing of oil sands in Alberta, Canada, where 1 Mt of solids material is moved each day [72]. One ﬁnal detail to be noted is that magnesium silicate minerals, such as olivine and serpentine, typically contain signiﬁcant amounts (in the order of 5–20 wt%) of iron oxides that can turn out to be valuable byproducts when produced in amounts too large to be overlooked by the iron- and steelmaking industries. 14.4.1 Material Resources

Although several different elements can be carbonated, the alkaline earth metals – calcium and magnesium – have proven to be the most suitable, due to their abundance and insolubility in Nature. Whilst iron has also been suggested for carbonation, it is a valuable mineral resource that is sought after for other purposes, and is therefore less suitable for large-scale carbonation implementations [16]. In addition to the abundant magnesium- and calcium-containing minerals, several industrial solid residues also exist that contain large amounts of Mg, Ca, and even Fe. Those minerals and other materials that have been investigated in relation to CO2 mineralization are listed in Table 14.2 [18]. Currently, the most investigated mineral resources are olivine, serpentine and wollastonite, although recently basalt has attracted increased interest [62, 74, 75]. In order to compare (on a theoretical basis) the CO2-binding capacity of a mineral source, Goff and Lackner [76] introduced the concept of RCO2. This gives the theoretical mass amount of a given mineral necessary to convert a unit mass of CO2 into mineral carbonate; thus, the lower the RCO2-value, the less mineral is required for carbonation. Typical values range from 1.8 to 3 ton mineral per ton CO2, depending of course heavily on the degree of carbonation conversion. In Ref. [62], RCO2 values are listed to range (given as ton mineral per ton CO2) from 1.6 for forsterite, 2.1 for serpentine/crysotile, 2.4 for basalt, 2.6 for wollastonite, to 6.3 for arothite (calcium aluminum silicate). Despite the amounts of Mg and Ca sources being vast, as mineral carbonation research continues to expand, then the need for a detailed worldwide evaluation of the amounts of suitable mineral deposits becomes more important [16]. A recently conducted evaluation of mineral reserves in the United States concluded that mineral resources are unlikely to be a limiting factor when industrial-scale mineral sequestration is considered [77]. It was estimated that there is enough

14.4 Mineral Carbonation for Carbon Capture and Storage (CCS) Minerals and other materials that are considered suitable for carbonation.

Table 14.2

Mineral

Composition

Other material

Composition

Basalt

b

Brucite

Mg(OH)2

Calcium silicate

CaSiO3, Ca2SiO4

Caustic lime

CaO

Enstatite

MgSiO3

(K, Na, Ca)1.2–2.0 (Fe , Al, Fe2+, Mg)4.0 x [Si7–7.6Al1–1.4O20](OH)4·nH2O

Forsterite

Mg2SiO4

FeTiO3

Hydromagnesite

(MgCO3)4·Mg(OH)2·4 H2O

Carbonated serpentinite

Limestone

CaCO3

Eclogite CaAl2Si2O8

Feldspar

Mg2SiO4

Forsterite Glauconite

a

Ilmenite Listwanite

e

3+

Magnetite

Fe3O4

Magnesia

MgO

Olivine

(Mg, Fe)2SiO4d

Magnesite

MgCO3

Opokaa

mainly CaCO3, SiO2, hematite and muscovite

Merwinite

Ca3Mg(SiO4)2

CaMgSi2O6 + (Fe, Al)

Nesquehonite

MgCO3·3 H2O

Serpentine

Mg3Si2O5(OH)4

Slaked lime

Ca(OH)2

Serpentinite

c

Talc

Mg3Si4O10(OH)2

Wollastonite

CaSiO3

Pyroxene f

a) b) c) d) e) f)

(Ca,Na)2(Mg,Al)(Si,Al)3O7

Not suitable for mineral carbonization due to high content of carbonates [73]. Basalt (%): for example SiO2: 49.2, TiO2: 1.8, Al2O3: 15.7, Fe2O3: 3.8, FeO: 7.1, MnO: 0.2, MgO: 6.7, CaO: 9.5, Na2O: 2.9, K2O: 1.1, P2O5: 0.4. Finnish serpentinite (Hitura) 83 wt% serpentine and 17 wt% magnetite (Fe3O4). In one case as Mg1.82Fe0.18SiO4. Studying reactions leading to the formation of listwanite. Further categorized into different mineral types: Chrysotile, lizardite, and antigorite.

material to sequester the total US (current level) CO2 emissions, which can be estimated at ∼7 Gt per year, for more than 500 years. Earlier [78], it was reported that an olivine-containing rock in Oman (350 × 40 × 5 km, 30% pure) would be able to ﬁx all CO2 that might be produced from the combustion of all carbon present on Earth; this principle is illustrated graphically in Figure 14.4. 14.4.2 Direct (Single-Step) Process Routes 14.4.2.1 Gas–Solid Processes Mineral carbonation was ﬁrst mentioned as a CO2-binding concept for CCS by Seifritz in 1990 [80]. A few years later, the concept of binding CO2 in calcium and

367

368

14 Fixation of Carbon Dioxide into Inorganic Carbonates

Figure 14.4 Estimated storage times and capacities for various CO2 sequestration methods. From Ref. [2] after Ref. [79].

magnesium carbonate minerals was further investigated in the US by Dunsmore [81], while later still this process, which is also known as “enhanced natural weathering,” was investigated in more detail by Lackner and coworkers at Los Alamos National Laboratory (LANL) [76, 82]. Natural silicate minerals such as olivine, serpentine and wollastonite, as well as basalt rock were identiﬁed as the most suitable raw materials, being both abundant and cheap. Since then, research relating to mineral carbonation has accelerated and become divided into two main CO2-binding approaches: (i) the direct method, where carbonation of the mineral takes place in a single process step; and (ii) the indirect method, where calcium or magnesium is ﬁrst extracted from the mineral and subsequently carbonated. These methods aim primarily at ex situ processing in a dedicated processing plant (as opposed to in situ carbonation by the injection of CO2 into geological formations). The ﬁrst experiments reported from the US showed a negligible carbonation of magnesium silicates at low CO2 pressures with 50–100 μm size particles; a pressure of 34 MPa eventually gave a carbonation conversion of 25–30% at 823 K after 2 h for a serpentinite [83, 84]. Later studies involved the carbonation of MgO and Mg(OH)2, to be extracted from magnesium silicates (i.e., indirect carbonation; see below). In Finland, a direct carbonation test with a serpentinite from Lapland in a pressurized thermogravimetric apparatus (PTGA) showed no conversion [85]. More recently, tests were conducted with a serpentinite mine tailing sample (∼83% serpentine, ∼14% magnetite Fe3O4, 3% others; size <75 μm) from the Hitura nickel mine in central Finland with (sub- and) supercritical CO2; direct carbonation

14.4 Mineral Carbonation for Carbon Capture and Storage (CCS)

during 3.5 h after heat-up and pressurization to 5.7–30 MPa, 423–473 K resulted in <2% carbonation. In a very recent study, the carbonation of Mg-silicates at 673–773 K, 100 MPa pressure with a (supercritical) “CO2-rich ﬂuid”, with or without NaCl present was reported to give between 3% and 57% carbonation conversion after 4 h. The CO2 source was oxalic acid, and the efﬁciency increased according to the ranking orthopyroxene, crystotile, olivine [86]. 14.4.2.2 Aqueous Solution Processes The complexity, unfavorable (energy) economics and overall CO2 balance of the magnesium silicate carbonation route using MgCl2 as intermediate (see Section 14.4.3.2) resulted in the development of a simpler, direct carbonation process using aqueous solutions. Shortly after the end of the 1990s, a research group in the US reported conversion rates of 65% in 1 h, and 80% within 30 min, as the result of the careful control of solution chemistry, heat treatment, and ﬁner grinding. Currently, this method is considered to be the most successful route for mineral carbonation using an aqueous solution of 0.64 M NaHCO3 and 1 M NaCl at 15 MPa and 458 K for olivine, 428 K for heat-treated serpentine, or 4 MPa and 473 K for wollastonite. The costs for CO2 sequestration would be 54, 78, and 64 US$ per ton CO2, respectively, for these three mineral feedstocks. For serpentine, a thermal treatment (at an optimal temperature of 888 – 903 K) was found to be more efﬁcient than mechanical treatment (grinding) [31, 87, 88], and in an assessment Herzog [89] concluded that the reported carbonation rates were obtained only at a high energy cost (“a 20% energy penalty for a coal-ﬁred power plant”). Although such cost levels would render this method too expensive at this point (see also Ref. [63]), this route still receives attention and improvements are being reported. Very recently, almost twice the extent of carbonation was reported by using an aqueous solution containing 5.5 M KHCO3, aimed at reducing the mineral pretreatment efforts and costs [90]. Another recent point of attention has been that the process costs have been overestimated, partly due to unrealistic calculations of energy efﬁciency and energy costs. The costs of the process heat input were signiﬁcantly overestimated when charged in the same way as power input, giving a false impression of the overall process economics. It is important to make the distinction between power and heat and the temperature of the various heat streams of the process. For heating purposes, it is not necessary – and even unwise – to use electric power, since it is enough to use heat of a sufﬁciently high temperature, which should result in a much lower cost. At a power or steel plant, this would mean using process waste heat instead of electricity [18, 91]. 14.4.3 Indirect (Multistep) Process Routes 14.4.3.1 Gas–Solid Processes Following efforts to directly carbonate magnesium silicate, studies conducted in the US during the late 1990s shifted to MgO extraction from magnesium silicates

369

370

14 Fixation of Carbon Dioxide into Inorganic Carbonates

and the effect of water (e.g., released from Mg(OH)2) on carbonation. The carbonation of Mg(OH)2 was soon found to be signiﬁcantly faster than that of MgO; using a 20 μm Mg(OH)2 powder, a conversion of 90% was obtained after 30 min at 838 K, 5.2 MPa [83, 84, 92]. Butt et al. [93] provided an analysis of Mg(OH)2 decomposition in helium and in CO2, at atmospheric pressure (∼0.075 MPa), using a combination of thermogravimetric analysis (TGA), X-ray diffraction (XRD) and electron microscopy. The presence of CO2 was found to reduce the rate of water release from Mg(OH)2 over the temperature range where MgCO3 was thermodynamically stable. With 50–100 μm particles, a maximum conversion rate was reported at 648 K, after 12 h in pure CO2, to yield ∼16.7 wt% carbonate. Other groups in the US later reported the details of studies on Mg(OH)2 decomposition [94] and the effect of pressure on its carbonation [95], by examining how hydroxylation and rehydroxylation interact with carbonation. In this case, increasing pressures were found to slow the dehydroxylation of Mg(OH)2, generating a smaller number of reactive MgO sites for carbonation. In Finland, investigations into stepwise carbonation processes initially involved the use of Finnish serpentine and (calcined) magnesium hydroxide powder (∼20 μm, 99% pure) in a PTGA with humid CO2, at 0.1 to 3.5 MPa pressure. The materials were ﬁrst heated to 1273 K, and then cooled in (mainly) CO2/N2 mixtures. Tests conducted at elevated pressures did not produce the expected increase in MgO carbonation rate, as the ﬁnal carbonation conversions fell from 5.6% at 0.1 MPa to 2.9% at 3.5 MPa in 99%/1% CO2/H2O. Hence, the decision was taken to proceed with Mg(OH)2 under test conditions where MgO formation was thermodynamically unfavorable [85, 96]. In addition to the PTGA tests, a series of investigations was conducted with Mg(OH)2 (Dead Sea Periclase, 150–710 μm, 97% pure) and MgO in an atmospheric bubbling ﬂuidized bed (FB) (batch) reactor. Above and below the calcination temperature, a mixture of sand and (calcined) Mg(OH)2 was ﬂuidized in pure CO2 over a period of 4–11 h. The product of carbonate material, which had built up on the reacting particles and eventually had slowed the chemical conversion, could be removed from the particles as ﬁnes by attrition and abrasion. These ﬁnes were entrained from the reactor with the exit gas ﬂow, facilitating their removal from the reactor. The fact that they had a considerably higher MgCO3 content (8.1 wt%) compared to the remaining material in the bed (4.4 wt%) [91, 99] suggested that ﬂuidized-bed reactors (FBRs) might be the most suitable for large-scale processing, as noted earlier by Lackner et al. [97]. PTGA tests on the carbonation of Mg(OH)2 (97%, 75–125 μm) continued in Finland, with tests at pressures up to 4.5 MPa and temperatures corresponding to 0.9·Teq(K), where Teq is the maximum temperature at which MgCO3 is stable for a given CO2 (partial) pressure; this provides a driving force –ΔG/T for the chemistry. Results acquired from a total nine tests conﬁrmed that Mg(OH)2 could be carbonated faster than MgO. Although the pressurization allowed for the use of higher temperatures, it was not easy to identify the corresponding optimal temperature. For a given pressure, the carbonation rate was seen to decrease with temperature as a result of thermodynamic limitations. Results based on the wet

Carbonation efficiency (%)

14.4 Mineral Carbonation for Carbon Capture and Storage (CCS)

60 50 40 30 20 10 0

4.5 643 733 4 3.5 768 783 798 813 0.11.2 Pressure (MPa) Temperature (K)

(a) Figure 14.5 (a) Results of conversion of Mg(OH)2 (75–125 μm) to MgCO3 after 6 h for various temperature/pressure combinations in 99%/1% CO2/H2O [98]; (b) Progress as

(b) calculated reaction times for full carbonation of Mg(OH)2 under a chemical kinetics control regime [18].

chemical analysis of samples before and after the tests were conducted are shown in Figure 14.5a [98]. An XRD analysis of the products showed indeed a signiﬁcant MgCO3 content in the ﬁnal material. The conversion at 4.0 MPa and 4.5 MPa was considerably faster than at 3.5 MPa at similar or slightly higher temperatures, but was still slow compared to the wet processes used in the US (see below), despite the progress made (see Figure 14.5b). Subsequently, unreacted shrinking core (USC) modeling was used to determine the rate-limiting steps and rate parameters and showed that, after ∼5% carbonation, the product layer diffusion rapidly became rate-determining [98]. This, again illustrated that it would be beneﬁcial to operate a gas–solid carbonation process in an FB reactor, where the particle collisions would cause product layer removal by attrition, abrasion, and fragmentation mechanisms. The most recent PTGA test series was aimed at monitoring the carbonation of serpentinite rock from the Hitura nickel mine in Finland (see above, particle size 74–125 μm) in three steps: (i) thermal activation by heat-up to 903 K in air during 1 h; (ii) hydroxylation of MgO to Mg(OH)2 at 3.5 MPa, 653 K; and (iii) carbonation at 3.5 MPa, at 768 or 798 K. This showed that, while the release of water from the mineral was fast and straightforward, the subsequent hydration of MgO to Mg(OH)2 using pressurized steam was surprisingly slow. The suggested reason for this was that the heat-up stage did not produce MgO but rather converted serpentine into olivine, water, and quartz. However, when the dehydroxylation/ hydroxylation of a Mg(OH)2 (Dead Sea Periclase, 97% pure, 75–125 μm) sample was studied separately, the conversion of MgO to Mg(OH)2 in pressurized steam was seen to be extremely slow. Besides faster carbonation kinetics for Mg(OH)2 compared to MgO, the energy economy of a three-stage process with Mg(OH)2 production from MgO, followed by carbonation, was also better

371

372

14 Fixation of Carbon Dioxide into Inorganic Carbonates

Figure 14.6 Schematic layout of the high-pressure, high-temperature ﬂuidized bed reactor with preheater set-up, at Åbo Akademi University.

than that of a two-stage process based on MgO carbonation [98]. Although the hydroxylation of MgO presents another challenge, this could be avoided by producing Mg(OH)2 directly from a silicate mineral. Current studies of the carbonation of magnesium silicates in Finland are focused primarily on the use of gas–solid FB reactors for processing via magnesium oxide and magnesium hydroxide intermediates, at carbonation temperatures and pressures up to 873 K, 10 MPa (allowing for supercritical conditions for CO2), supported by earlier experiments using PTGA and extensive thermodynamic analyses [2, 98, 99]. A test facility for this, which is shown schematically in Figure 14.6, is currently being taken into use. Part of these investigations involves the carbonation of Mg(OH)2 (or MgO), while at the same time developing methods (“wet” or “dry”) for producing this from serpentinite rock. This is to be followed by a second phase in which these synthetic MgO or Mg(OH)2 materials are carbonated [100]. Raw material can be found in vast amounts in Finland: besides hoisted material, the serpentinites in Eastern Finland (the Outokumpu-Kainuu ultramaﬁc rock belt) alone could provide sufﬁcient material for 200–300 years of CCS processing (based on a current CO2 emission excess in Finland of ∼11 Mt per year, with respect to the 1997 Kyoto Protocol commitments) [24]. 14.4.3.2 Aqueous Solution Processes The efforts to extract MgO from magnesium silicates before carbonation resulted in an evaluation of a process where hydrochloric acid (HCl in water) could be used to extract Mg from minerals, via the intermediates MgCl2·nH2O and Mg(OH)Cl, to be followed by gas–solid carbonation of Mg(OH)2 [17, 83]. Because of the com-

14.4 Mineral Carbonation for Carbon Capture and Storage (CCS)

plexity and the energy consumption of this process route, other US research groups have continued working on the direct route for mineral carbonation, based on systems using aqueous solutions. Due to the need for an extensive pretreatment of the feedstock mineral, those aqueous processes which use direct carbonation could in fact be considered as indirect. As an alternative to energy-intensive pretreatment, several groups worldwide have embarked on methods to either dissolve or leach the Mg or Ca from minerals or industrial byproducts and wastes, by using strong or weak acids (mineral or organic), alkali solutions, or ligands. [101–105]. An example of this is the process route developed by Kakizawa et al. [50] in Japan (see Section 14.3.3.2). One major problem with these routes is that chemicals are used that typically cannot be recovered and reused, and this leads to deteriorating process economics. For example, Teir and coworkers [54, 106] dissolved serpentinite from the Hitura nickel mine in both weak and strong aqueous solutions of common acids or bases at room temperature during 1 h. Subsequently, it was found that H2SO4 extracted most Mg, followed by HCl, HNO3, HCOOH, and CH3COOH. None of the acids extracted Mg selectively, but extracted also some Fe and Si. Carbonate precipitation experiments were conducted with two solutions prepared from serpentinite in strong solutions of HNO3 and HCl, respectively; after ﬁltration and drying, the remaining salts were dissolved in water and the iron oxides precipitated. The magnesium-rich solutions (mainly dissolved magnesium nitrates or chlorides) were used to precipitate magnesium carbonates by exposing each solution to a CO2 ﬂow, while regulating the pH by the addition of aqueous NaOH (this was necessary to produce precipitates). Although, as a result quite pure hydromagnesite, Mg5(CO3)4(OH)2·4 H2O was produced, the costs of the pH-regulating agent NaOH and for make-up acid made this route economically nonviable. Nonetheless, the general understanding of magnesium and calcium carbonation reactions has improved signiﬁcantly (see also the studies by Hänchen et al. [107–110] on the relative importance of process parameters such as temperature, CO2 pressure and particle size distribution). Studies involving a three-step process of olivine carbonation, involving: (i) dissolution of olivine; (ii) precipitation of magnesite; and (iii) precipitation of silica in an aqueous solution, were recently reported from Norway [69], where the process proceeds without chemical additives at 10–15 MPa and 403–523 K. No reaction rates were reported, however. At present, many studies are ongoing to identify a means of enhancing the carbonation chemistry of magnesium silicates in aqueous systems, using weak acids and additives that will improve silica dissolution, such as citrates, oxalates, and EDTA [105, 111]. In this case, a near-complete recovery and reuse, thereby minimizing the losses of such chemicals, will be essential for viable process economics. Likewise, there is much to improve with regards to the reaction rates and/ or times. Recently, basalt materials, with MgO + CaO-contents typically of the order of 20 wt% or less, but widely available, have been considered for CO2 mineralization [74, 75] (see also the Hellisheidi project on Iceland [62]).

373

374

14 Fixation of Carbon Dioxide into Inorganic Carbonates

14.5 Other Carbonate Production Processes and Applications 14.5.1 Carbonation of Brines

Brine is a saline-based solution that is formed as a waste product during oil or natural gas extraction (over 75 million m3 is produced each year in the US alone [112]), and as such it can be found stored in vast quantities in above-ground storage tanks. This large amount and relatively high concentration of metals that are capable of forming carbonates (mainly Ca and Mg), provides a carbonation process option for CO2 storage. Yet, despite the ability of brine to form carbonates, any industrial-scale operation is currently limited by the slow reaction kinetics. Raising the pH of the brine would speed up the carbonation process, but uncertainties concerning the parameters (brine composition, temperature, pressure and pH) require further investigation [113]. In another recent study, Soong et al. [114] investigated the possibility of using ﬂy ash to raise the pH of brine, thereby allowing for the precipitation of carbonates from the solution. These initial investigations conﬁrmed the feasibility of the concept, with 0.546 mol l−1 of CO2 being sequestered in 2 h during a one-stage approach via ﬂue-gas desulfurization (FGD) ﬂy ash. These experiments were performed in an autoclave reactor with an initial CO2 pressure of 0.136 MPa at 293 K. A conceptual model of the process is given in Figure 14.7 (marked by the solid arrows; the dotted arrows show other alternative brine/ash CO2 sequestration routes that have been investigated [114]). Other brines and mud suspensions from industrial processes that are currently being considered for carbonation include aqueous red mud ﬂows [27] and mixtures of bauxite and saline waste water [115] from aluminum production. Here, the CO2 ﬁxation capacities are of the order of 5–10 g per 100 g red mud as a result of 15 wt% Na2O + CaO (dry), while a 90%/10% (v/v) bauxite waste/brine mixture can take up more than 9.5 g CO2 per liter of liquid. Worldwide, a total of 200 Mt bauxite residues are stored, mainly in ponds. Fly ash

Raw brine from well

Adjust AdjustpH pH

Caustic

Separator Separator

CO2

Reactor Reactor T, T,PP

Recycled fly ash

Separator Separator

Reacted brine

Carbonates Inject Inject underground underground

Figure 14.7

CO2 sequestration process via brine and ﬂy ash [114].

Acknowledgments

14.5.2 Straightforward Carbonation

The simple spreading of, for example, olivine on land where acidity is a problem would simultaneously increase the pH of the soil (i.e., improve soil quality) and capture CO2 from the surrounding air within a relatively short time frame (∼30 years). This simple approach to CCS has been suggested by Schuiling and Krijgsman [11], who have emphasized that this method, although simple, should initially be applied with caution so as to conﬁrm the impact of spreading large amounts of rock material on the ground. The amount of CO2 that could be sequestered in this way is principally limited by the available/suitable surface area and the theoretical binding capacity (see Equation 14.2). It should be noted that this reaction is heavily dependent on the rainfall, soil type, CO2 pressure, temperature, and type of rock, all of which limit its applicability. Another alternative, yet simple, approach to CO2 sequestration might be carbonation in underground cavities, such as caves. Schuiling [116] has discussed the possibility of sequestering CO2 by ﬁlling, for example, an open-cast mine with olivine-containing rock material, and then injecting the CO2. The beneﬁts of such a solution are that no expensive reactor equipment would be required, and that the reaction kinetics would not be of major importance. In addition, the heat generated by the reaction between CO2 (and H2O) and olivine could be recovered by placing heat exchangers in the olivine. According to Schuiling, this may represent a breakthrough to the solution of the greenhouse gas problem [116]. However, it should be noted that, unless the kinetics are fast enough, the system of olivine and CO2 would reach thermal equilibrium with the surrounding rock material, and heat recovery would not be possible. A signiﬁcant carbonation could also take place in mine tailings without any intervention, although quantifying the amount of CO2 trapped within the tailings would be difﬁcult (see Ref. [117]). Nonetheless, this method could be used at mine sites with tailings rich in serpentine, thus providing the mining operation with “CO2 credits.”

Acknowledgments

R.Z. wishes to acknowledge the many years of project cooperation and discussions with Dr Sebastian Teir (currently at VTT Technical Research Centre of Finland) and Mrs Sanni Eloneva at Helsinki University of Technology, Finland. Note: The data reported in Ref. [18] forms the basis of the present text. Sebastian Teir is also acknowledged for artwork in Ref. [18], used here as Figures 14.1, 14.2, 14.3 and 14.7.

375

376

14 Fixation of Carbon Dioxide into Inorganic Carbonates

References 1 USGS, US Geological Survey (2009) Minerals Information, http://minerals. usgs.gov/minerals/ (accessed 6 March 2009). 2 Zevenhoven, R., Eloneva, S., and Teir, S. (2006) Catal. Today, 115, 73–79. 3 Roskill (2008) Roskill Information Services, 7th edn (see also URL: http:// www.roskill.com/reports.html) (accessed 6 March 2009). 4 Moulijn, J.A., Makkee, M., and van Diepen, A. (2001) Chemical Process Technology, John Wiley & Sons, Ltd, UK. 5 Olper, M. and Fracchia, P. (1988) US Patent 4,769,116. 6 Brownlow, A.H. (1996) Geochemistry, 2nd edn, Prentice-Hall, USA, pp. 355–369. 7 Tarbuck, E.J. and Lutgens, F.K. (2005) Earth – An introduction to physical geology, Pearson Education International, USA. 8 Rau, G.H. and Caldeira, K. (1999) Energy Convers. Management, 40, 1803–1813. 9 Rau, G.H., Knauss, K.G., Langer, W.H., and Caldeira, K. (2007) Energy, 32, 1471–1477. 10 Stolaroff, J.K., Lowry, G.V., and Keith, D.W. (2005) Energy Convers. Management, 46, 687–699. 11 Schuiling, R.D. and Krijgsman, P. (2006) Clim. Change, 74, 349–354. 12 Gao, Q., Tao, Z., Huang, X., Nan, L., Yu, K., and Wang, Z. (2009) Geomorphology, 106, 324–332. 13 Jha, P.K., Tiwari, J., Singh, U.K., Kumar, M., and Subramanian, V. (2009) Chem. Geol., 264, 364–374. 14 Salminen, J. and Prausnitz, J. (2007) Developments and Applications in Solubility (ed. T.M. Letcher), RSC Publishing, UK, pp. 337–349. 15 Huijgen, W.J.J. and Comans, R.N.J. (2003) Report ECN-C–03-016, Energy Research Centre of The Netherlands. 16 Huijgen, W.J.J. and Comans, R.N.J. (2005) Report ECN-C–05-022, Energy Research Centre of The Netherlands.

17 Newall, P.S., Clarke, S.J., Haywood, H.M., Scholes, H., Clarke, N.R., and King, P.A. (2000) Report PH3/17, IEA Greenhouse Gas R&D Prog. 18 Sipilä, J., Teir, S., and Zevenhoven, R. (2008) Report VT 2008-1, Åbo Akademi University, Heat Engineering Laboratory. 19 Fernández Bertos, M., Simons, S.J.R., Hills, C.D., and Carey, P.J. (2004) J. Hazard. Mater., 112, 193–205. 20 Li, X., Bertos, M.F., Hills, C.D., Carey, P.J., and Simon, S. (2007) Waste Management, 27, 1200–1206. 21 Teir, S., Eloneva, S., Fogelholm, C.-J., and Zevenhoven, R. (2007) Energy, 32, 528–539. 22 Huijgen, W.J.J. (2007) Carbon dioxide sequestration by mineral carbonation, PhD (Eng) thesis, University of Wageningen, The Netherlands. 23 Huijgen, W.J.J., Witkamp, G.J., and Comans, R.N.J. (2005) Environ. Sci. Technol., 39, 9676–9682. 24 Teir, S. (2008) Fixation of carbon dioxide by producing carbonations from minerals and steelmaking slags, PhD (Eng) thesis, Helsinki University of Technology, Espoo, Finland. 25 Eloneva, S. (2008) Co-utilization of CO2 and steelmaking slags for precipitated calcium carbonate production, Lic. tech. thesis, Helsinki University of Technology, Espoo, Finland. 26 Lekakh, S.N., Rawlins, C.H., Robertson, D.G.C., Richards, V.L., and Peaslee, K.D. (2008) Metall. Mater. Trans. B., 39, 125–134. 27 Bonenfant, D., Kharoune, L., Sauvé, S., Hausler, R., Niquette, P., and Mimeault, M. (2008) Ind. Eng. Chem. Res., 47, 7610–7616. 28 Yamasaki, A., Iizuka, A., Kakizawa, M., Katsuyama, Y., Nakagawa, M., Fujii, M., et al. (2006) Fifth Annual Conference on Carbon Capture and Sequestration, Alexandria, Virginia, USA. 29 Huntzinger, D.N., Gierke, J.S., Sutter, L.L., Kawatra, S.K., and Eisele, T.C. (2009) J. Hazard. Mater., 168, 31–37.

References 30 Huntzinger, D.N., Gierke, J.S., Sutter, L.L., Kawatra, S.K., and Eisele, T.C. (2009) Environ. Sci. Technol., 43, 1986–1992. 31 Gerdemann, S.J., O’Connor, W.K., Dahlin, D.C., Penner, L.R., and Rush, H. (2007) Environ. Sci. Technol., 41, 2587–2593. 32 Baciocchi, R., Polettini, A., Pomi, R., Prigiobbe, V., VonZedtwitz, V.N., and Steinfeld, A. (2006) 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway. 33 Baciocchi, R., Polettini, A., Pomi, R., Prigiobbe, V., Von Zedwitz, V.N., and Steinfeld, A. (2006) Energy Fuels, 20, 1933–1940. 34 Rendek, E., Ducom, G., and Germain, P. (2007) Waste Management, 27, 1403–1407. 35 Rendek, E., Ducom, G., and Germain, P. (2006) J. Hazard. Mater., 128, 73–79. 36 Kuusik, R., Uibu, M., and Trikkel, A. (2006) 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway. 37 Uibu, M., Uus, M., and Kuusik, R. (2009) J. Environ. Management, 90, 1253–1260. 38 Uibu, M., and Kuusik, R. (2007) The 3rd International Green Energy Conference, Västerås, Sweden. 39 Eloneva, S., Teir, S., Salminen, J., Fogelholm, C., and Zevenhoven, R. (2008) Ind. Eng. Chem. Res., 47, 7104–7111. 40 Iizuka, A., Fujii, M., Yamasaki, A., and Yanagisawa, Y. (2004) Ind. Eng. Chem. Res., 43, 7880–7887. 41 Costa, G., Baciocchi, R., Polettini, A., Pomi, R., Hills, C.D., and Carey, P.J. (2007) Environ. Monit. Assess., 135, 55–75. 42 Baciocchi, R., Costa, G., Polettini, A., Pomi, R., and Prigiobbe, V. (2008) 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), Washington (DC). 43 Abanades, J.C., Rubin, E.S., and Anthony, E.J. (2004) Ind. Eng. Chem. Res., 43, 3462–3466. 44 Chen, Z.Y., O’Connor, W.K., and Gerdemann, S.J. (2006) Environ. Prog., 25, 161–166.

45 Domingo, C., Loste, E., GómezMorales, G.J., and Fraile, J. (2006) J. Supercrit. Fluids, 36, 202–215. 46 Domingo, C., García-Carmona, J., Loste, E., Fanovich, A., Fraile, J., and Gómez-Morales, J. (2004) J. Cryst. Growth, 271, 268–273. 47 Feng, B., Yong, A.K., and An, H. (2007) Mater. Sci. Eng.: A, 445-446, 170–179. 48 Filgueiras, A.V., Lavilla, I., and Bendicho, C. (2002) J. Environ. Monit., 4, 823–857. 49 Rimstidt, J.D., Balog, A., and Webb, J. (1998) Geochim. Cosmochim. Acta, 62, 1851–1863. 50 Kakizawa, M., Yamasaki, A., and Yanagisawa, Y. (2001) Energy, 26, 341–354. 51 Teir, S., Eloneva, S., and Zevenhoven, R. (2005) PRES’05, Giardini di Naxos / Messina, Italy. 52 Teir, S., Eloneva, S., and Zevenhoven, R. (2005) Energy Convers. Manag., 46, 2954–2979. 53 Eloneva, S., Teir, S., Savolahti, J., and Zevenhoven, R. (2007) ECOS ‘2007, Padova, Italy, pp. 1389–1396. 54 Teir, S., Revitzer, H., Eloneva, S., Fogelholm, C.-J., and Zevenhoven, R. (2007) Int. J. Miner. Process., 83, 36–46. 55 Teir, S., Eloneva, S., Fogelholm, C., and Zevenhoven, R. (2007) IGEC-III, Västerås, Sweden. 56 Yogo, K., Teng, Y., Yashima, T., and Yamade, K. (2004) 7th International Conference on Greenhouse Gas Control Technologies (GHGT-7), Vancouver, BC, Canada. 57 Katsuyama, Y., Yamasaki, A., Iizuka, A., Fujii, M., Kumagai, K., and Yanagisawa, Y. (2005) Environ. Prog., 24, 162–170. 58 Geerlings, J.J.C. and Wesker, E. (2007) Patent PCT/EP 2006/068691. 59 Gorset, O., Johansen, H., Kihle, J., Munz, I.A., and Raaheim, A. (2007) Patent WO/2007/069902. 60 Yogo, K., Eikou, T., and Tateaki, Y. (2005) JP Patent 2005097072. 61 Kodama, S., Nishimoto, T., Yogo, K., and Yamada, K. (2006) 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway.

377

378

14 Fixation of Carbon Dioxide into Inorganic Carbonates 62 Oelkers, E.H., Gislason, S.R., and Matter, J. (2008) J. Elem., 4, 333–337. 63 Intergovernmemtal Panel on Clinate Change (2005) Carbon Dioxide Capture and Storage (eds B. Metz, O. Davidson, H.C. de Coninck, M. Loos, and L.A. Meyer), Special report Prepared by Working Group III of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. 64 Teir, S., Eloneva, S., Fogelholm, C.-J., and Zevenhoven, R. (2006) Energy Convers. Management, 47, 3059–3068. 65 Koljonen, T., Siikavirta, H., Zevenhoven, R., and Savolainen, I. (2004) Energy, 29, 1521–1527. 66 Lee, J.K. and Dong, W.K. (2005) 22nd Annual International Pittsburgh Coal Conference, Pittsburgh, PA, USA. 67 Kavaliauskaite, I., Uibu, M., Teir, S., Kuusik, R., Zevenhoven, R., and Denafas, G. (2005) ECOSUD2005/ 5th International Conference on Ecosystems and Sustainable Development, Cadiz, Spain. 68 Stasiulaitiene, I., Denafas, G., and Sliaupa, S. (2007) 32nd International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, Florida, USA. 69 Munz, I.A., Kihle, J., Brandvoll, Ö., Machenbach, I., Carey, J.W., and Haug, T.A. (2008) 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), Washington (DC). 70 Munz, I.A., Korneliussen, A., Gorset, O., Johansen, H., Kihle, J., and Flaathen, T. (2006) 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway. 71 International Energy Agency (2008) CO2 capture and storage – A key carbon abatement option, International Energy Agency, Paris 2008. 72 Kunzig, R. (2009) National Geographic, 215 (3), 38–59. See also: http://ngm. nationalgeographic.com/2009/03/ canadian-oil-sands/kunzig-text (accessed 7 April 2009). 73 Kavaliauskaite, I., and Denafas, G. (2006) 8th International Conference on

74

75

76 77

78 79 80 81 82

83

84

85

86

87

88

Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway. McGrail, B.P., Schaef, H.T., Ho, A.M., Chien, Y., Dooley, J.J., and Davidson, C.L. (2006) J. Geophys. Res., 111, B12201. Schaef, H.T., McGrail, B.P., and Owen, A.T. (2008) 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), Washington (DC). Goff, F. and Lackner, K.S. (1998) Environ. Geosci., 5, 89–101. Krevor, S.C., Graves, C.R., Van Gosen, B.S., and McCafferty, A.E. (2009) Mapping the mineral resource base for mineral carbon-dioxide sequestration in the conterminous United States, U.S. Geological Survey Digital Data Series 414, http://pubs.usgs.gov/ds/414/ (accessed 9 March 2009). Lackner, K.S. and Ziock, H. (2000) Modern Power Systems, 20, 31–32. Lackner, K.S. (2003) Science, 300, 1677–1678. Seifritz, W. (1990) Nature, 345, 486. Dunsmore, H.E. (1992) Energy Convers. Management, 33, 565–572. Lackner, K.S., Wendt, C.H., Butt, D.P., Joyce, E.L., and Sharp, D.H. (1995) Energy, 20, 1153–1170. Lackner, K.S., Butt, D.P., and Wendt, C.H. (1997) 22nd International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, Florida, USA, pp. 419–430. Lackner, K.S., Butt, D.P., and Wendt, C.H. (1997) Energy Convers. Management, 38, 259–264. Zevenhoven, R., Kohlmann, J., and Mukherjee, A.B. (2002) 27th International Conference on Coal Utilization and Fuel Systems, Clearwater, Florida, USA, pp. 743–754. Dufaud, F., Martinez, I., and Shilobreeva, S. (2009) Chem. Geol., 262, 344–352. O’Connor, W.K., Dahlin, D.C., Rush, S.J., Gerdemann, S.J., and Penner, L.R. (2004) DOE/ARC-2004-2028. O’Connor, W.K., Dahlin, D.C., Rush, G.E., Gerdemann, S.J., Penner, L.R., and Nilsen, R.P. (2005) DOE/ ARC-TR-04-002, Albany Research Center.

References 89 Herzog, H. (2002) Report 14 March, 2002, MIT Laboratory for Energy and the Environment. 90 Plasynski, S., Lang, D.A., and Chizmeshya, A. (2008) Project 314, U.S. DOE Project Facts. 91 Zevenhoven, R., Sipilä, J., and Teir, S. (2008) 2nd International Conference on Accelerated Carbonation for Environmental and Materials Engineering, ACEME-08, Rome (Italy), pp. 45–54. 92 Butt, D.P., Benjamin, A., Holesinger, T., Park, Y.S., Lackner, K.S., and Wendt, C.H. (1997) Global Warming International Conference, Columbia University, New York. 93 Butt, D.P., Lackner, K.S., Wendt, C.H., Conzone, S.D., Kung, H., and Lu, Y. (1996) J. Am. Ceram. Soc., 7, 1892–1898. 94 Chizmeshya, A.V.G., McKelvy, M.J., Sharma, R., Carpenter, R.W., and Bearat, H. (2003) Mater. Chem. Phys., 77, 416–425. 95 Béarat, H., McKelvy, M.J., Chizmeshya, A.V.G., Sharma, R., and Carpenter, R.W. (2002) J. Am. Ceram. Soc., 4, 742–748. 96 Zevenhoven, R. and Kavaliauskaite, I. (2004) Int. J. Thermodyn., 7, 22–31. 97 Lackner, K.S., Butt, D.P., Wendt, C.H., Goff, F., and Guthrie, G. (1997) Report LA-UR-97-2094, LANL. 98 Zevenhoven, R., Eloneva, S., and Teir, S. (2006) 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway. 99 Zevenhoven, R. and Teir, S. (2004) Third Annual Conference on Carbon Capture and Sequestration, Alexandria (VA), USA, p. 217. 100 Fagerlund, J., Teir, S., Nduagu, E., and Zevenhoven, R. (2008) 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), Washington (DC). 101 Blencoe, J.G., Palmer, D.A., Anovitz, L.M., and Beard, J.S. (2004) Patent WO 2004/094043. 102 Park, A.A. and Fan, L. (2004) Chem. Eng. Sci., 59, 5241–5247. 103 Maroto-Valer, M.M., Zhang, Y., Kuchta, M.E., Andresén, J.M., and Fauth, D.J. (2005) US Patent 2005/0002847.

104 Maroto-Valer, M.M., Fauth, D.J., Kuchta, M.E., Zhang, Y., and Andrésen, J.M. (2005) Fuel Process. Technol., 86, 1627–1645. 105 Krevor, S.C. and Lackner, K.S. (2008) 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), Washington (DC). 106 Teir, S., Kuusik, R., Fogelholm, C.-J., and Zevenhoven, R. (2007) Int. J. Miner. Proc., 85, 1–15. 107 Hänchen, M., Prigiobbe, V., Baciocchi, R., and Mazzotti, M. (2008) Chem. Eng. Sci., 63, 1012–1028. 108 Hänchen, M., Krevor, S., Mazzotti, M., and Lackner, K.S. (2007) Chem. Eng. Sci., 62, 6412–6422. 109 Hänchen, M., Prigiobbe, V., Storti, G., and Mazzotti, M. (2007) 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8), Trondheim, Norway. 110 Hänchen, M., Prigiobbe, V., Storti, G., Seward, T.M., and Mazzotti, M. (2006) Geochim. Cosmochim. Acta, 70, 4403–4416. 111 Park, A.A. and Fan, L. (2008) 2nd International Conference on Accelerated Carbonation for Environmental and Materials Engineering, ACEME-08, Rome, Italy. 112 Jones, A., Glass, C., Reddy, T.K., Maroto-Valer, J.M., Andrésen, J.M., and Schobert, H.H. (2001) ACS Div. Fuel Chem., 46 (1), 321. 113 Druckenmiller, M.L. and Maroto-Valer, M.M. (2005) Fuel Process. Technol., 86, 1599–1614. 114 Soong, Y., Fauth, D.L., Howard, B.H., Jones, J.R., Harrison, D.K., and Goodman, A.L. (2006) Energy Convers. Management, 47, 1676–1685. 115 Dilmore, R., Lu, P., Allen, D., Soong, Y., Hedges, S., and Fu, J.K. (2008) Energy Fuels, 22, 343–353. 116 Schuiling, R.D. (2006) Macroengineering – A Challenge for the Future (eds V. Badescu, R.B. Cathcart, and R.D. Schuiling), Springer, Dordrecht, The Netherlands. 117 Wilson, S.A., Raudsepp, M., and Dipple, G.M. (2006) Am. Mineral., 91, 1331–1341.

379

381

Index a accelerated weathering of limestone (AWL) 355 – plant 355 acetanilide–CO2 complex 101 acetic acid 282 – route 361 acetyl-CoA 37ff. – reductive acetyl-CoA pathway (Wood–Ljungdahl pathway) 39 acetyl-CoA/propionyl-CoA carboxylase 40ff. acidiﬁcation phase 347 acrylic acid 22 acylating agent 131 – transfer of carbamate group 131 adipic acid 18 alcohol – CO2 reduction 299 – transfer of carbamate group 134 alcoholysis – urea 174 aldehyde 282 algae 341f. aliphatic hydrocarbon – reforming with CO2 275 aliphatic primary amine 127 alkali carbamate 127 alkali metal phenoxide 95 alkali metal phenoxide–CO2 complex 93 alkali phenoxide – CO2-solvated 93 alkali phenyl carbonate (PhOCO2M) 93 alkane 89 – carboxylation 89 alkene 89 – carboxylation 89 alkenyl-boronic ester 107 alkyl halide 127 – transfer of carbamate group 126

N-alkyl propargylamine – carboxylative cyclization 148 3-alkyl-5-hydroxyoxazin-2-one 141 3-alkyl-4-hydroxyoxazolidin-2-one 141 alkylammonium N-alkylcarbamate (RNH3)O2CNHR 132 alkylammonium carbamate 122ff. alkylboronic ester – copper-catalyzed carboxylation 107 alkylene carbonate – cyclic 198 – ﬁve-membered 205 – six-membered 205 N-alkylprop-2-ynylamine – carboxylation 148 alkyne 89 – 1-alkyne 158 – carboxylation 89 allylamine 144 alpha-aryl-propionic acid 319 alumina 148 aluminum – Ni/γ-Al2O3 catalyst 251 – (salen)AlX 221 – (5,10,15,20-tetraphenylporphyrinato) aluminum(III) acetate 140 – tetraphenylporphyrinato(aluminum) chloride 217 aluminum–magnesium complex – (R2Al(μ-NSiMe3)(μ-OSiMe3)Mg(thf)2 (μ-O2C))3 65 aluminum–porphyrin complex 140, 183, 217 amide 131 amidine 124ff. amino alcohol – ring closure 137 ammonia-oxidizing bacteria – anaerobic 40

Carbon Dioxide as Chemical Feedstock. Edited by Michele Aresta Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32475-0

382

Index anode – sacriﬁcial 318ff. Aquifex 39 Aquiﬁcales 49 – hydrogen-oxidizing microaerobic 39 Archaea – methanogenic 40 Archaeoglobus 40 Archaeum – anaerobic autotrophic hyperthermophilic 44 Arnon–Buchanan Cycle 37 aromatic carboxylic acid – Kolbe–Schmitt synthesis 90ff. aromatic ketone – electrocarboxylation 324 aromatics 89 – carboxylation 89 arothite 366 artiﬁcial photosynthetic cycle 7 aryl-acetylene 111 N-arylcarbamate 124 arylcarboxylic acid 107 associative attachment 271 atom economy (AE) factor 173 atom/ion/molecule reaction 270 ATP 38ff. attachment 270 autotrophic carbon ﬁxation – diversity 46 – pathway 33ff. autotrophs 33 aziridine – 1-alkyl-2-aryl-substituted 136 – copolymer from carbon dioxide 245 – derivative 214

b bacteria – acetogenic 40 – anaerobic ammonia-oxidizing 40 – photosynthetic 347 – sulfate-reducing 40 barium – BaTiO3 286 benzoic acid derivative 107 1-benzyl-1,4-dihydronicotinamide (BNAA) 294 benzyltrimethylammonium hydroxide 130 beta-zeolite 131 [6,6′-bi(5,7-dimethyl-1,4,8,11tetraazacyclotetradecane)]-dinickel(II) triﬂate 296

bimetallic catalytic system 104ff. bio-oil 339ff. biodiesel 339ff. – energy 349 biofuel – ﬁrst-generation 339 – new generation 339 – second-generation 340 – third-generation 341 biogas 343ff. bioglycerol 345 bioinspired reaction 82 biological ﬁxation 10 biomass 7, 335ff. – aquatic 335ff. – cultivated 339 – residual 338 – technology cascade 338 – terrestrial 337ff. bioreﬁnery concept 347ff. biotin enzyme 122 biotin-dependent carboxylase 82 (2,6-bis[(di-t-butylphosphino)methyl]phenyl) ligand 104f. 1,3-bis(2,6-diisopropylphenyl)imidazolium2-carboxylate 99 bis(oxazolidin-2-one) derivative 141 1,3-bis(silyl) carbodiimide 150 2,2-bis(triﬂuoromethyl)-4,5-diﬂuoro-1,3dioxole (PDD) 25 bisphenol-A-polycarbonate (BPA-PC) 172ff., 244 boronic ester/copper(I) salt/CsF catalytic system 107 Botryococcus braunii 341 bridged complex 67 – in-situ synthesis 67 brines – carbonation 374 N-(2-bromoacyl)oxazolidin-2-one – chiral 324 α-bromoethylbenzene 324 (E)- and (Z)-β-bromostyrene 327 N-bromosuccinimide (NBS) 188 [(n-Bu)4N](PF6) 304 n-Bu3P/DBAD (di-tert-butyl azodicarboxylate) 137 butadiene – Pd(0)-catalyzed telomerization with CO2 112 tert-butyl hydroperoxide (TBHP) 188 1-butyl-3-methylimidazolium benzene sulfonate [BMIm][PhSO3] 192

Index 1-butyl-3-methylimidazolium tetraﬂuoroborate [BMIm][BF4] 131, 188, 326 1-butyl-3-methylimidazolium-2-carboxylate 98, 192 1-butyl-3-methylimidazolium-4-carboxylate 98 1-butyl-3-methylimidazolium-5-carboxylate 98 1-(4-iso-butylphenyl) ethyl chloride 324 2-(4-iso-butylphenyl) propionic acid (S-ibuprofen) 324 1-(4-iso-butylphenyl)-1-chloroethane 320

c C–C double bond – transfer of carbamate group 142 C–C triple bond – transfer to carbamate group 145 C–O bond cleavage 70 cadmium – colloidal particles of CdS 310 caesium – Cs2CO3 134 – CsOH/BMImCl 156 calcium – CaO 201 – CaO–ZrO2 201 calcium aluminum silicate 366 calcium carbonate 354ff. – ground (GCC) 354 – precipitated (PCC) 354ff. – production 359 calcium silicate 361 Calvin–Benson–Bassham (CBB) cycle 33ff., 47 carbamate ester 125 carbamate group 121 – transfer to acylating agent 131 – transfer to alcohol 134 – transfer to alkyl halide 126 – transfer to C–C double bond 142 – transfer to C–C triple bond 145 – transfer to epoxide 138 carbamate salt ionic liquid 243 carbamation 133 carbamic acid – synthesis 121ff. carbamoylhydrazine 153 carbon analyzer 266 carbon capture and storage (CCS) 353ff. – mineral carbonation 364 carbon cycle – natural 336

carbon deposition – in-situ study 249ff. carbon dioxide 1ff. – accumulation in the atmosphere 1 – acetic acid 282 – activated state 69 – acyclic carbonate 172ff. – additive to CO for the synthesis of methanol 267 – aldehyde 282 – alkali phenoxide 93 – carbon monoxide production 272 – catalyst for amine methoxycarbonylation 132 – complex 64ff. – conditions for using 8 – coordinated 55ff. – coordination chemistry 55ff. – coordination modes in polynuclear metal–CO2 complexes 58 – copolymer from aziridine 245 – copolymerization of epoxide 215ff., 228 – copolymerization of oxetane 213, 228ff. – copolymerization of oxirane 228 – cyclic ketal 191 – cyclourethanization of 1,2-amino alcohol 136 – dense 15ff., 134 – diol 193 – dissociation 272 – electrochemical carboxylation 317ff. – electrochemical reduction 296ff. – electroreduction in protic media 327 – emission 1 – enzymatic catalysis 18 – feedstock 172 – ﬁxation into inorganic carbonate 353ff. – glycerol 195 – higher hydrocarbons 279 – hydrocarbon synthesis 277 – hydrogen production 272 – hydrogenation 280 – in chemical synthesis 11 – inert solvent for chemical synthesis 15ff. – inorganication 12 – δ-lactone 112 – liquid 15 – metal 56 – methanol 280f. – natural ﬁxation in carbonate 355 – nonthermal plasma approach 267ff. – oxidant in selective processes 267 – oxidative coupling reaction of CH4 277

383

384

Index – oxygenates synthesis 280ff. – Pd-catalyzed ﬁxation 183 – Pd(0)-catalyzed telomerization of butadiene 112 – photochemical reduction 292ff. – photoelectrochemical reduction 305ff. – plasma utilization 284 – polymer 213ff. – polymer synthesis 20 – potential of utilization 9 – price 8 – radical anion 298ff. – reaction intermediates in CO2 utilization processes 75 – reaction with electrophile 72 – reaction with nucleophile 73 – reactivity of CO2 complexes 69f. – reduction 276, 291ff. – reforming 60, 249ff., 273ff. – solvent medium for chemical process 15f. – source 8 – stable complex 64 – supercritical (scCO2) 15ff., 123ff., 136ff., 154ff., 215, 242, 267, 369 – technological application 10 – technologies to reduce accumulation in the atmosphere 4f. – utilization 6ff. – utilization of terrestrial and aquatic biomass 335ff. carbon dioxide complex – synthesis and structure 59 carbon disulﬁde 214 – copolymerization of epoxide 244 – copolymerization of thiorane 244 carbon ﬁxation – autotrophic 33ff. – chemical aspect 50 – mechanism 34 carbon formation 250ff. – H2O effect 262 – kinetic study on partial pressure of CH4 and CO in CO2 reforming 260 – TEM analysis after CO2 reforming of methane 259 – thermodynamic analysis 252ff. carbon monoxide – carbon dioxide 272 – CO2 reduction 298 – deoxygenation 250 – disproportionation 250ff. – metallocarboxylate 293 – production 302

– thermodynamic analysis of carbon formation 252 carbon monoxide dehydrogenase (CODHase) 69 carbon-concentrating mechanism 47 carbonate – acyclic 200 – acyclic organic 170ff. – cyclic (CC) 116, 181ff. – ﬁve-membered alkylene 205 – inorganic 353 – linear 169ff. – natural ﬁxation of CO2 355 – organic cyclic 181ff. – process route 357 – six-membered alkylene 205 carbonation – aqueous solution process 369ff. – brines 374 – carbon capture and storage (CCS) 364 – direct 357ff. – gas–solid 358ff. – indirect 357ff., 369 – mineral 364 – multistep process route 369 – single-step process route 358ff. – solid residue 362 – two-step aqueous 362 carbonic anhydrase 83 carboxylase 34ff. – biotin-dependent 82 carboxylation – N-alkylprop-2-ynylamine 148 – C–H-acidic compound 99 – CO2 317ff. – electrochemical 317ff. – epoxide 182ff. – main group element-promoted 89ff. – oxidative 186 – oleﬁn 186 – trans 102 – transition metal-promoted 89ff. carboxylative cyclization 192 – N-alkyl propargylamine 148 carboxylic acid 102 – α,β-unsaturated 110 – organometallic derivative in the synthesis 102 carboxylic anhydride 131 catalyst recycling 241 catalytic process 81 catalytic system – bimetallic 104ff.

Index cathode – carbon-based cathode with Pt nanoparticles 329 – copper 328 cellulose – dissolution in ionic liquid 340 Cenarchaeum symbiosum 43 ceria 193, 284 cerium – CeO2/γ-Al2O3 276 – H3PW12O40/Ce0.1Ti0.9O2 177 chaetomellic acid A anhydride 109 chain polymerization – synthesis of ﬂuoropolymer 22 charge transfer 271 – homogeneous catalyst (HCTC) 322 chemical utilization 10ff. chemosynthesis 34 chloroacetonitrile 320 Chlorobium 39 – tepidum 37 Chloroﬂexaceae 44 Chloroﬂexus 44 – aggregans 41 – aurantiacus 41ff. chloroﬂuorocarbon (CFC) 23 chloroplast 33 chromium – carbon dioxide complex 61 – Cr(III) porphyrin complex 218 – Cr(PIBsalen) catalyst 241 – (salan)CrX catalyst 227 – salen complex 244 – (salen)Cr(III)(DMAP) 136 – (salen)CrCl catalyst 219ff. – (salen)CrCl/n-Bu4NN3 system 239f. – [trans-(salen)Cr(N3)2][n-Bu4N] 230 – [trans-(salen)Cr(N3)2][PPN] 230 – (salen)CrN3/PPNN3 catalyst 237 – [trans-(salen)CrX2][n-Bu4N] 231 – [trans-(salen)CrX2][PPN] 231 – (salen)CrX complex 222ff. – (tmtaa)CrCl catalyst 224 cinnamic acid – derivative 107 – (E)- and (Z)-cinnamic acid 327 citric acid cycle – oxidative (Krebs cycle) 37 – reductive (Arnon–Buchanan Cycle) 37 Cladophora 341 climate change power (CCP) 7 CO dehydrogenase/acetyl-CoA synthase 40 cobalamin 39

cobalt 191 – (C3H5N2)4(SiMo11CoO38(CO2))·4 H2O 66 – carbon dioxide complex 61 – Co(II) tetra-azamacrocyclic complex 296 – Co-aminophosphane complex CoCl(NO)2[PhP(OCH2CH2)2NH] 123 – [Co(bpy)n]+ 294 – Co(Pr-salen)K(CO2) 64 – Co(salen) complex 242, 321 – Co(tetraphenylporphyrin) 321 – Na[Co(salen)(N3)2] complex 225 – porphyrin complex 226, 321 – (salen)CoX 221 – Zn-Co double metal cyanide (DMC) 244 Codium fragile 341 collisional detachment 271 coordination chemistry – carbon dioxide 55ff. copolymer – isolation 241 copolymerization 219 – carbon disulﬁde and a thiorane 244 – epoxide and carbon disulﬁde 244 – epoxide and CO2 215ff., 228 – episulﬁde 244 – oxetane and CO2 213, 228 – oxirane and CO2 228 copper 191 – boronic ester/copper(I) salt/CsF catalytic system 107 – carbon dioxide complex 61ff. – carboxylation of alkylboronic ester 107 – cathode 328 – Cu(I) carbamate 126 – Cu(I) halide 147 – Cu/B 107 – Cu-coated glassy carbon electrode 301 – Cu1.5PMo12O40 177 – (NHC)Cu(OBpin) 79 corona discharge 268ff. Crassulacean acid metabolism 47 Crenarchaea 49 Crenarchaeota 42f. – anaerobic 39 – (micro)aerobic 43 crown-ether (CE) 127, 189 – 18-crown-6-ether 129 cyanoacetic acid 111, 326 cyclam (1,4,8,11-tetra-azacyclo tetradecane) 296 cyclic alkylene carbonate 198 cyclic carbamate 137ff. cyclic carbonate (CC) 116, 181ff. – halogenated carbonate 190

385

386

Index – halohydrin 189 – propargylic alcohol 191 – scCO2 185 – scCO2 and ionic liquid 185 – synthesis 204 cyclic ester 110 cyclic ketal 191 β-cyclodextrin (βCD) 182, 324 cyclohexane-2,6-dicarboxylic acid 100 cyclohexene oxide monomer 235 N-cyclohexyl-N′,N′,N″,N″-tetraethylguanidine (CyTEG) 125 cyclourethanization of 1,2-amino alcohol 136 CyTMG 130, 143

d DBAD (di-tert-butyl azodicarboxylate) 137 DEADC/PR3 137 decomposition 270 – methane 250 Deltaproteobacteria 39f. dense gas 12 density functional theory (DFT) calculation 59 deoxygenation of carbon monoxide 250 Desulfobacter 38 Desulfobacterium sp. 40 Desulfotomaculum sp. 40 Desulfurococcales 44 detachment 270 N,N-dialkylcarbamate 131 [(N,N-dialkylcarbamoyl)oxy]ethylene 145 1,3-dialkylimidazolium tetraﬂuoroborate 102 1,3-dialkylimidazolium tetraphenylborate 102 1,3-dialkylimidazolium-2-carboxylate – synthesis 97 diaryl-imidazolylidene 111 1,5-diazabicyclo[4.3.0]non-5-ene (DBN) 193 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) 100f., 130ff., 143, 188ff. dibenzylamine (Bz2NH) 123 dibromo(1,4,8,11-tetraazocyclotetradecane)nickel(II) 136 dibutylmethoxymethylcarbamato stannane 175 dicarbamoylhydrazine 153 dicarboxylate/4-hydroxybutyrate cycle 44ff. dicyclohexylcarbodiimide (DCC) 176 dielectric-barrier discharge (DBD) 268, 284 Diels–Alder reaction 19 diethyl carbonate (DEC) 176, 204f.

diethyl peroxydicarbonate (DEPDC) 24 diethylazodicarboxylate/PPh3 176 β-[(N,N-diethylcarbamoyl)oxy]styrene 145 1,3-diisopropyl-4,5-dimethyl-imidazolium-2carboxylate 98 1,3-dimesitylimidazolium-2-carboxylate 98 2,2-dimethoxypropane (DMP) 176f. dimethyl carbonate (DMC) 170ff., 204 – direct carbonation 181 – DMC-DPC 180 – phosgene route 171 – transesteriﬁcation 181 – urea methanolysis 181 dimethyl ether (DME) 176f. 4,6-(dimethylamino)pyridine (DMAP) 136, 226 1,8-dimethylaminonaphthalene (DMAN) 157 N,N-dimethylaminoquinoline (DMAQ) 226 1,3-dimethylimidazolium-2-carboxylate 98ff., 192 3,3-dimethyloxetane 233 diol – carbon dioxide 193 – urea 194 1,2-diol 138 – monocarbamate 138 diphenyl carbodiimide (DPC) 100f. diphenyl carbonate (DPC) 172, 204 1,3-diphenyl urea (DPU) 100f. 4R-(diphenylmethyl)-oxazolidin-2-one 326 (DIPHOS)Pd(Cl)(nortricyclo urethane) 142f. dispersion polymerization 21 – surfactant 21 dissociation 270 dissociation attachment 270 dissociation ionization 270

e electrocarboxylation 111, 318ff. – aromatic ketone 324 – organic halide 319 electrochemical reduction 296ff. electron acceptor – terminal 33 electron/molecular reaction 270 electroreduction – CO2 327 end-on complex 65 energy – biodiesel 349 energy density 6 enzyme 305, 335

Index enzymatic catalysis – dense carbon dioxide 18 epihalohydrin 189 episulﬁde 214, 244 epoxide – alicyclic 213 – aliphatic 213 – carboxylation 182ff. – copolymerization 215 – ionic liquid 184 – transfer of carbamate group 138 Epsilonproteobacteria 39 erythrocarine 109 EtBuNH 243 ethanol 339 ethene carbonate (EC) 169, 199 1-ethoxyethyl carbamate 142 ethylene/CO2 coupling reaction – Ni(0)-promoted 112 2-ethylheptanoic acid 114 2-ethylidene-6-hepten-5-olide 112 2-ethylidene-7-morpholinohept-5-enoic acid 115 Euryarchaeota 40 excitation 270

f fatty acid 343ff. fatty acid methyl ester (FAME) 345 fayalite 356 fenopren 320 ferredoxin 45 ferric oxide 356 Fischer–Tropsch synthesis 279 ﬂue-gas desulfurization (FGD) 374 9-ﬂuorene-carboxylic acid 101 ﬂuoropolymer – synthesis 22 formate dehydrogenase 1 enzyme – tungsten-containing 305 formic acid – CO2 reduction 298 forsterite 356 fossil carbon 348 Friedel–Craft reaction – acid-catalyzed 19 fruit vegetable garden (FGV) 347 fuel – algal biomass-derived 344

g gallium – p-GaP electrode 306f. gas discharge 286

gas-to-liquid (GTL) – approach 12 – conversion 267 gel-permeation chromatography (GPC) 235 glass transition temperature 22 glassy carbon electrode – Cu-coated 301 gliding arc discharge 269 glow discharge 268 glyceraldehyde 3-phosphate 35ff. glycerol 195, 346 – aqueous 345 – carbon dioxide 195 – urea 195 glycerolysis 198 – urea 198 glyoxylate 40 gold – Au(III) 192 green chemical synthesis 19 green chemistry 170ff. green engineering 172ff. Grignard reagent 103

h halohydrin 189 hematite 356 heptanoic acid 108 heterogeneous photochemical semiconductor system 309 heterogeneous reaction 271 hexaalkylphosphorous triamide P(NR2)3 155 hexaﬂuoropropylene (HFP) 25 trans-hexahydro-3-methylbenzoxazolidinone 136 hexamethylene diisocyanate (HDI) 149 hexylzinc iodide-lithium chloride complex 108 homogeneous charge-transfer catalyst (HCTC) 322f. hydroamination 115 hydrocarbon 277ff. – CO2 reduction 300 – hydrogenation of CO2/CO to C2 hydrocarbon 278 – solvent 21 – synthesis 277 hydroformylation 16ff., 115 hydrogen – CO2 272 hydrogenation 17f., 115 – C2 hydrocarbon 278 Hydrogenobacter 39 hydromagnesite 373

387

388

Index hydrosilylation 73 hydrotalcite 148 2-hydroxy-benzoic acid 90ff. 4-hydroxy-benzoic acid 90ff. 2-hydroxy-2-(4-i-butylphenyl) propionic acid 325 2-hydroxy-2-(6-methoxy-2-naphthyl) propionic acid 325 3-hydroxy-2-naphthoic acid 90 6-hydroxy-2-naphthoic acid 90 hydroxoalkylamine 200 4-hydroxybutyryl-CoA dehydratase 42f. trans-2-hydroxycyclohexyl N,N-disubstituted carbamate 138 3-hydroxypropionate/4-hydroxybutyrate cycle 42ff. 3-hydroxypropionate/malyl-CoA cycle 40ff. HZSM-5 zeolite 286

i ibuprofen 320ff. – S-ibuprofen 324 Ignicoccus 45 – hospitalis 44 imidazolidinone 199 2-imidazolium carboxylate 98 indium – p-InP electrode 307 iodoalkyloxazinanone 144 iodoalkyloxazolidinone 144 ion recombination 271 ionic liquid (IL) 136, 175, 189ff., 242, 326 – carboxylation of epoxide 184 – dissolution of cellulose 340 – scCO2 185 ionization 270 iron – carbon dioxide complex 61 – Fe(CO2)(depe)2 72 – Fe(CO2)(PMe3)4 80 – (Fe(dmpe)2(CO)H)(HCO3) 80 – Fe(PMe3)3(CO)(CO3) 80 – FeC(O)OR 73 – FeCl(CO2SnR3)(depe)2 73 – hematite (ferric oxide) 356 isocyanate RNCO 121ff., 148 isokinetic effect 133 isophorone diisocyanate (IPDI) 149

k 3-ketoglutaric acid 100 ketone – electrocarboxylation of aromatic ketone 324

Kolbe–Schmitt carboxylation 50, 97 Kolbe–Schmitt synthesis 90ff. Krebs cycle 37

l

δ-lactone 112ff. – synthesis 112 lanthanoid – Ln{N(SiMe3)2}3 150 lanthanum – carbon dioxide complex 61 – La2O3/γ-Al2O3 276ff. laser ablation 59 lignin 340ff. limestone 353 limonene oxide 217 lipid 345 – transesteriﬁcation 345 liqueﬁed natural gas (LNG) extraction 12 lithium phenoxide 95 low heating value (LHV) 344 low-temperature matrix isolation 59

m macroalgae 344 macrocyclic complex 296ff. macromolecule – rational design 21 magnesite 354 magnesium – carbonate 363 – hydroxide 370ff. – Mg–Al–CO3 hydrotalcite 202 – Mg5(CO3)4(OH)2·4 H2O 373 – Mg(N(SiMe3)2)(OSiMe3) 65 – MgO 201 – (R2Al(μ-NSiMe3)(μ-OSiMe3)Mg(thf)2 (μ-O2C))3 65 – silicate 356ff., 369 main group element-promoted carboxylation – organic substrate 89 malonyl-CoA 40ff. malyl-CoA/β-methylmalyl-CoA/ citramalyl-CoA lyase 42 mangenese – CpMn(η5-ArO) 74 – (CpMn(μ-OAr)(THF))2 75 – Mn6(O2CNEt2)12 132 mass control 8 material resource 357ff. mesoporous graphitic carbon nitride (mpg-C3N4) 122 mesoporous silica (MCM-41) 131

Index metal – carbon dioxide 56 – catalyst 215ff., 228ff. – CO2 complex intermediate 293 – coordination modes in polynuclear metal–CO2 complexes 58 – electrode 302 metal N-alkylcarbamate 131 metal formate 293 metal phthalocyanine complex 131 metal-containing enzyme-mediating complex 305 Metallosphaera sedula 44 metastable de-excitation 271 methane – acetic acid 282 – aldehyde 282 – CO2 reduction 300 – CO2 reforming 249ff. – complete oxidation 250 – dry-reforming 12 – decomposition 250ff. – higher hydrocarbons from CO2 279 – methanol 281 – oxidative coupling reaction with CO2 277 – partial oxidation 250 – reforming with CO2 273 – steam reforming 250 – synthesis gas production 249ff. – thermodynamic analysis of carbon formation 252ff. methanol 281 – CO2 as additive to CO 267 – CO2 hydrogenation 280 – CO2 reduction 299 Methanomicrobia 37 para-methoxy-benzoic acid 107 3-methoxymethyl-3-methyloxetane 233 p-methoxyphenol/DMAP system 136 methyl α-cyanophenylacetate 102 methylene diphenyl diisocyanate (MDI) 148 5-methylene-1,3-oxazolidin-2-one 148 4-methylene-2-oxazolidinone 147 5-methylene-2-oxazolidinone 148 – N-unsubstituted 148 methylmethacrylate 22 4-methyloxazol-2-one 147 4-methyloxazolidin-2-one 135 N-methylpyrrolidone (NMP) 157 methyltetrahydropterin 39 methylurethane 126 microalgae 344 microwave discharge 269

mineral – magnesium- and calcium-containing 366 Mitsunobu transformation 137 Mitsunobu zwitterion 153 molecular sieve – n-Bu2N-MCM-41 201 molecular weight distribution 217 molybdenum – trans-Mo(C2H4)2(PMe3)4 77 – trans-Mo(CO)(CO3) HN(CH2CH2PMe2)2(PMe3) 67 – Mo(CO3)(CO)2(PMe3)3 67 – trans-Mo(CO2)2HN(CH2CH2PMe2)2(PMe3) 67 – Mo(η2-CO2)(η3-P4)(dppe) 70 – fac-Mo(CO)(η3-P4=O)(dppe) 70 monocarbamates 138 – 1,2-diol 138 morpholine 115 MTO (methyltrioxorhenium) 188 MTO/UHP/Zn[EMIm]2 Br4/[BMIm]BF4 188

n N–CO2 bond – chemistry 121ff. NAD(P)H 35ff. Nannoclhoropsis sp. 341 naproxen 325 natural gas discharge 286 natural resource 353 neutral abstraction 271 neutral recombination 271 nickel – [6,6′-bi(5,7-dimethyl-1,4,8,11tetraazacyclotetradecane)]-dinickel(II) triﬂate 296 – carbon dioxide complex 61 – catalyst 252 – (Cy3P)2Ni(η2-C,O)–CH2=CO 75 – ((IPr)Ni)2(μ-CO)(μ-η2,η2-CO2) 67 – L2Ni(CO2)(C2H4) 77 – Ni(0)/acetylenes/CO2 catalytic system 112 – Ni(acac)2 108 – Ni/α-Al2O3 258 – Ni/γ-Al2O3 catalyst 251ff., 284 – Ni(bipy)(C2H4) complex 112 – Ni/CaO/Al2O3 catalyst 258 – Ni2(μ-CNMe)(CNMe)2(dppm)2 149 – Ni(CNR)4 149 – Ni(CO)2(CNR)2 149 – [Ni(cyclam)]2+ 296 – Ni(cyclam)Br 183

389

390

Index – Ni(1,5-cyclo-octadiene)2/phosphine catalytic system 110 – Ni/MgO 258 – Ni(PCy3)2 109 – Ni(PCy3)2(CO2) 11, 64, 84 – Ni/SiO2 catalyst 258 – Ni/Zn catalytic system 108 – NiBr2(cyclam) 136 – NiCl2(dppe) 321 – (PCy3)2Ni(CO4) 76 – (PNP)Ni 152 niobium – alkoxide 203 – Nb(IV) complex 183 – Nb(CO2)(C5H4Me)(CH2SiMe3) 64 – Nb(OCH3)5 178f. – NbO2 204 nitrile effect 113 Nitzschia sp. 341 nonsteroidal anti-inﬂammatory drug (NSAID) 319f. nonthermal plasma approach 267ff. – utilization of carbon dioxide 267ff. nonthermal plasma phenomena 268

o O transfer 70 oleﬁn epoxidation 18 olivine 356, 375 organic halide – electrocarboxylation 319 organolithium compound 103 organometallic derivative – synthesis of carboxylic acid 102 organotin compound 174 organotin iodide complex 229 Oscillochloris sp. 41 osmium – carbon complex 63 – cis-[Os(bpy)2(CO)H][PF6] 304 – Os4(μ-H)(μ-CO2)(thd)(CO)13 68 oxalate production 302 oxanickelacycle 112 oxazolidin-2-one 135f. oxazolidinone 199 oxetane – copolymerization of CO2 213, 228ff. – metal catalyst 228ff. – monomer 239 oxidation – aerobic 18 – methane 250 – partial oxidation of alcohol 18 – partial oxidation of methane 250

– selective oxidation of organic substrate 17 oxidative carboxylation – oleﬁn 186 oxidative citric acid cycle (Krebs cycle) 37 oxidative coupling reaction 76 – CH4 with CO2 277 oxirane – copolymerization of CO2 228 2-oxo-cyclohexane-1-carboxylate 102 2-oxo-cyclohexane-carboxylic acid 101 O-β-oxoalkyl-N,N-dialkyl carbamate 147 2-oxoglutarate synthase 37 β-oxopropylcarbamate 147 oxygen – oxidant 186 oxygenates synthesis 280ff.

p palladium 191 – (π-allyl)palladium 143, 183 – (1,5-cyclooctadiene)PdCl2 142 – (dicyclopentadiene)PdCl2 142 – (DIPHOS)Pd(Cl)(nortricyclo urethane) 142 – modiﬁed electrode 308 – (norbornadiene)PdCl2 142 – PCP–Pd(II) allyl complex 106 – (PCP)Pd–Me complex 105 – (PCP)Pd–OAc 105 – Pd(0)-catalyzed telomerization of butadiene with CO2 112 – Pd(0)/phosphine complex 143 – Pd(II) 192 – Pd(II)-complex 122 – Pd(acac)2 112 – Pd/γ-Al2O3 276ff. – Pd-COOMe 78 – [Pd(η3-2-Me-C3H4)(OAc)] 112 – Pd(O2CCH3)2 148 – Pd(PBu3)4 104 – [Pd(PCy3)2] 105 – Pd(PPh3)4 104 – [Pd(PPh3)2(p-benzoquinone)] 112 – [Pd(PPh3)4/PPh3] 183 – Pd/Sn system 104 – [Pd(triphosphine)L](BF4)2 304 – PdCl2(MeCN)2 157 – PdCl2(PPh3)2 321 PCP-pincer 104ff. penning dissociation 271 penning ionization 271 penta-alkylguanidine 130 – superbase 124

Index perﬂuorinated cyclic amine 23 perﬂuoro-vinyl-ether 25 perﬂuoroalkyl sulﬁde 23 perﬂuorocarbon 23 perﬂuoropolyether 24 perﬂuoropropyl-vinylether (PPVE) 25 peroxocarbonate LnM(CO4)Xm 84 pH swing process 363 phenol 50 1-(3-phenoxyphenyl)-1-chloroethane 320 phenylacetonitrile – carboxylation 102 5-phenyloxazolidin-2-one 135 phosgene 170 phosphazene 130 phosphine 191 phosphine imide reaction 156 phosphocarbamate P(NR2)(O2CNR2)2 127, 155 phosphoenolpyruvate (PEP) carboxylase 44 3-phosphoglycerate 35 phosphoribulokinase 37 photobioreactor (PBR) 348 photocatalyst 293ff. photochemical reduction 292ff. – homogeneous 292 photoelectrochemical reduction 305ff. photoelectrochemical semiconductor electrode system 305 photosensitizer 294ff. – ruthenium-based 294 photosynthesis 34 photosynthetic cycle – artiﬁcial 7 plant – terrestrial 341 plasma – interaction with catalyst 285 plasma utilization 284 – catalyst 284 platinum – nanoparticle on carbon-based cathode 329 – Pt/CeZrOx 258 – Pt(cod)Cl2 115 poly(ε-caprolactone) 27 poly(cyclohexene carbonate) 218, 235 poly(cyclohexylene carbonate) 228ff. poly(dithiocarbonate) 244 poly(ethylene carbonate) 215 poly(isobutylene) (PIB) 241 poly(D,L-lactide-co-glycolide) 27 poly(propylene carbonate) 215, 228ff. poly-pyrones 110

poly(TFE-co-PSEVPE) 25 poly(TFE-co-Vac) 25 poly(trithiocarbonate) 244 poly(trimethylene carbonate) (poly(TMC)) 228ff., 240 poly(urethane-amine) 214 poly(VDF-co-HFP) 26 poly(vinylidene ﬂuoride) (PVDF) 23 polycarbonate 213 polydimethylsiloxane monomethacrylate (PDMS-mMA) 24 polydispersity 217 polyether – macrocyclic 129 polymer – biodegradable 26 polymer synthesis 20 – supercritical carbon dioxide 20 polymerization – dispersion 21 – particle-forming chain free-radical 22 – precipitation 24 – ring opening (ROP) 27 – scCO2 26 – solubility 21 – step 26 – surfactant 21 – synthesis of biodegradable polymer 26 polytetraﬂuoroethylene (PTFE) 23 polythiocarbonate 213 polytrithiocarbonate 214 polyurethane 213, 245 potash 354 potassium feldspar 356 potassium phenoxide 99 potassium phenoxide–CO2 complex 96 precipitated calcium carbonate (PCC) production process 361 pressurized thermogravimetric apparatus (PTGA) 368ff. propane-1,3-diol 205 propargylic alcohol – cyclic carbonate 191 propene carbonate (PC) 169 propionyl-CoA 40ff. propionyl-CoA synthase 42 3-propyl 2-oxo-1,3-oxazoline 147 propylene oxide monomer 237 proton exchange membrane fuel cell (PEMFC) technology 328 2-pyrones 110 pyrrolidin-2-one 130 pyruvate : ferredoxin oxidoreductase (pyruvate synthase) 44

391

392

Index

q quenching – direct 292 – reductive 292

r recombination 270 reductant – sacriﬁcial 292ff. reduction 79, 291ff. – aqueous solution at metal electrode 297 – carbon dioxide 276, 291ff. – electrochemical 296ff. – electroreduction in protic media 327 – metal complex 303 – nonaqueous solution at metal electrode 302 – photochemical 292ff. – photoelectrochemical 305ff. reductive citric acid cycle (Arnon–Buchanan Cycle) 37 resource 357 RF discharge 269 rhenium 295ff. – (CO)5Re2(CO2)2Re(CO)4)2 64 – photocatalyst 294f. – fac-Re(bpy)(CO)3Cl 303 – {Re(bpy)(CO)3[P(OEt)3]}SbF6 295 – [Re(bpy)(CO)3(POiPr)3]+ 295 – Re(CO)3(bpy)Cl 294f. – [Re(CO)3(v-bpy)Cl] 308 – [Re(dmb)(CO)3]2 complex 295 rhodium – catalyst 16f., 115 – Rh(I) catalyst 107 – Rh(III) 192 – Rh/Al2O3 251 – Rh/B 107 – cis-[Rh(III)(bpy)2Cl2]2+ 304 – mer-Rh(C6H5)(H)Cl(PMe3)3 81 – Rh(diars)2Cl(CO2) 64 – RhCl(CO)(PMe3)2 82 – RhCl(η2-O2)(P)3 84 – RhClP3 186 Rhodospirillum rubrum 347 ribulose 1,5-bisphosphate 35 ribulose-1,5-bisphosphate carboxylase/ oxygenase (RubisCO) 34f., 82, 122, 335 ring opening polymerization (ROP) 27, 232 – cyclic thiocarbonate 244 – trimethylene carbonate 232 Roseiﬂexus spp. 41

rubidium – Rb2CO3 134 RubisCO-like protein (RLP) 37 ruthenium 191, 293f. – carbon complex 61 – catalyst 158, 251 – (η4-1,5-cyclooctadiene)(η6-1,3,5cyclooctatriene)ruthenium 148 – [Ph(CH2)nPPh2]Ru(η3-CH2=C(Me)CH2)2 145 – Ru(III) 192 – [Ru(bpy)3]2+ 294 – [Ru(bpy)2(CO)2]2+ 304 – [Ru(bpy)2(CO)CH3CN] 293 – [Ru(bpy)2(CO)Cl]+ 304 – [Ru(bpy)2(CO)H]+ 293f. – Ru3(CO)12 147 – Ru(COD)(COT) 148 – [Ru(phen)3]2+ 294 – cis-Ru(phen)2(CO)(CHO)(PF6) 68 – (Ru(phen)2(CO)(CO2)Ru(bpy)2(CO))(PF6)2 68 – (Ru(phen)2(CO)(CO2)Ru(phen)2(CO))(PF6)2 68 – Ru(tpy)(4,4′-X2bpy)(OCHO)+ 75 – [RuCl2(η6-arene)]2 145 – RuCl3·xH2O 145 – [RuCl2(norbornadiene)]n 145ff. – trans-[RuCl2{P(OC2H5)3}4] 145 – RuCl2(PMe2Ph)(p-cymene) 147 – cis/trans-(RuX2(MeCN)4) 73 – mer-(RuX3(MeCN)3) 73 – RuX2PR3(η6-arene) 145

s salan (N,N′-disubstituted bis(aminophenoxide)) 221 salen – H2salen (N,N′-bis(salicylidene)ethane-1,2-diamine) 321 salen–metal complex 242 salicylic acid 91ff. saturated calomel electrode (SCE) 297 scCO2 123ff., 136ff., 154ff., 215, 242, 267, 369 – cyclic carbonate (CC) 185 – ionic liquid 185 – polymer synthesis 20 – solubility 22 Schizodetrium sp. 341 selective energy transfer (SET) model 133 semiconductor – metal-coated colloid and powder 310 – unmodiﬁed colloid and powder 309

Index semiconductor electrode – homogenous solution catalysts 308 semiconductor electrode interface – modiﬁed 307 – Pd-modiﬁed 308 – unmodiﬁed 306 semiconductor system – heterogeneous photochemical 309 – reduction 305 side-on complex 67 silent discharge 268 silicate – weathering 353ff. silicon – coated p-SiC powder 310 – H2N(CH2)2NH(CH2)3Si(OMe)3 123ff. – H2N(CH2)3Si(OEt)3 123ff. – H2N(CH2)3Si(OMe)3 123ff. – H2NC(O)NH(CH2)2NH(CH2)3Si(OMe)3) 123 – MeO(O)CNH(CH2)2NH(CH2)3Si(OMe)3 133 – MeO(O)CNH(CH2)3Si(OMe)3 133 – MeO(O)CNH(CH2)3Si(OMe)x(OEt)3−x 133 – (MeO)3Si(CH2)3NH2+CH2CH2NHCO2− 123 – NaN(SiMe3)2 150 – p-Si 307 silyl isocyanate 150 silylcarbamate 154f. N-silylamines RNH(SiMe3) 154 sodium carbonate 353 sodium hydroxide (caustic soda) 353 sodium phenoxide–CO2 complex 95 sodium phosphotungstate 188 solubility – scCO2 21 solvent medium – chemical processes 15ff. – inert 15ff. sputtering 271 steam reforming of methane 250 step polymerization – synthesis of biodegradable polymer 26 styrene 22 succinyl-CoA 42ff. Sulfolobales 44 – autotrophic 42 sulfonyl-vinyl ﬂuoride 25 sulfur bacteria – anaerobic phototrophic green 39 sulfur oxidizer – microaerobic 39 super base 193 supercritical ﬂuid (SCF) 15ff.

supercritical gas 15 superoxide 130 switchable polarity solvent (SPS) 243 syngas 345ff. synthesis 271 synthesis gas production 249ff. Syntrophobacter fumaroxidans 305

t tamoxifen 109 tapered element oscillation microbalance (TEOM) 252ff. – carbon formation in CO2 reforming of methane 256 – measurement 264 TBAB (Bu4N)Br 130, 188 telomerization – Pd(0)-catalyzed 112 temperature-programmed oxidation (TPO) analysis 251ff. – carbon formation in CO2 reforming 257 tetraethylammonium perchlorate (TEAP) 131 tetraethylurea 157 tetraﬂuoroethylene 22 tetraphenylantimony halide 136 (5,10,15,20-tetraphenylporphinato) aluminum(III) acetate 140 Thermoproteales 44 Thermoproteus 45 – neutrophilus 44 – sp. 39 thiocarbonate – cyclic 244 – ROP 244 thiorane 244 – copolymerization of carbon disulﬁde 244 tin – bis[N,N-bis(dimethylphenysilyl)amido]tin(II) 152 – bis[N-trimethylsilyl-N-2,6diisopropylphenylamido]tin(II) 152 – n-Bu2SnO 195 – n-Bu2Sn(OCH3)2 179, 195f. – n-Bu2Sn[OCH(CH3)2]2 179 – (CH3)2Sn(OCH3)2 179 – dibutyltin(IV) 174 – R2Sn(OCH3)2 178 – Sn(OMe)2 195 titanium – carbon complex 61 – Cu-loaded TiO2 powder 310 – H3PW12O40/CexTi1−xO2 177 – metal-coated TiO2 powder and colloid 310

393

394

Index – Ti(IV) 177 – (Ti(IV), U(V)) arylimido-complex 150 – WO3-doped TiO2 powder 310 titanosilicate molecular sieve 131 toluene diisocyanate (TDI) 149 transesteriﬁcation 200ff. – lipid 345 transition metal catalyst 228 transition-metal complex 61ff., 294ff. – Ni(CO2)(PCy3)2 11 transition-metal-promoted carboxylation – organic substrate 89 transmission electron microscopy (TEM) 266 – analysis after CO2 reforming of methane 259 tri-reforming process 249 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) 194 triazolinone 153 tributyl(allyll)tin 104 triethanolamine (TEOA) 294 triglyceride 345 trimethylene carbonate – ring opening polymerization 232 trimethylsilyl carbamate 154 triphenylphosphine/diethylazodicarboxylate (Ph3P/DEAD or DEADC) 135, 156 triphenylstibine oxide 136, 157 – Ph3SbO/P4S10 system 158 Triton-B (benzyltrimethylammonium hydroxide) 130 tungsten – carbon dioxide complex 61 – formate dehydrogenase 1 enzyme 305 – H3PW12O40/CexTi1−xO2 177 turnover frequency (TOF) 217

u uranium 71f. – (AdArO)3tacnUIII complex 65ff. – U(III) complex 80 urea 154ff. – alcoholysis 174 – diol 194 – glycerol 195 – glycerolysis 198 – (RNH)C(O)(NR′R″) 157

urea hydrogen peroxide (UHP) 188 urethane 131ff. utilization – aquatic biomass 335ff. – terrestrial biomass 337ff.

v v-bpy (4-vinyl-4′-methyl-2,2′-bipyridine) 308 vanadium – V2O5 204 – VO(OiPr)3 203 vinyl acetate (VAc) 25 vinyl carbamate 145 vinylidene ﬂuoride (VDF) 22 vinyloxazolidinone 143 N-vinylpyrrolidone 22 volatile organic compound (VOC) 23

w water gas-shift (WGS) reaction (WGSR) 55, 250, 347 weathering – natural and artiﬁcial 353ff. – silicate 353ff. wollastonite 361 Wood–Ljungdahl pathway 39

z zeolite 284 – beta-zeolite 131 – HZSM-5 286 zinc – glutarate 215 – highly ﬂuorinated carboxylate 218 – K2CO3–KI/ZnO 202 – KI/ZnO 202 – metalloenzyme 83 – phenoxide 100 – Zn-Co double metal cyanide (DMC) 244 – ZnS colloidal suspension 310 zirconia 177f., 193 zirconium – CaO–ZrO2 201 – H3PW12O40/ZrO2 177 – ZrO2 177f., 193 zirconium–borane complex – LZr(CH2Ph)(B(C6F5)4) 80